You are on page 1of 28

Netherlands Journal of Geosciences

http://journals.cambridge.org/NJG
Additional services for Netherlands

Journal of Geosciences:

Email alerts: Click here


Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

Towards an improved geological interpretation of airborne


electromagnetic data: a case study from the Cuxhaven tunnel valley and
its Neogene host sediments (northwest Germany)
D. Steinmetz, J. Winsemann, C. Brandes, B. Siemon, A. Ullmann, H. Wiederhold and U. Meyer
Netherlands Journal of Geosciences / FirstView Article / January 2015, pp 1 - 27
DOI: 10.1017/njg.2014.39, Published online: 30 December 2014

Link to this article: http://journals.cambridge.org/abstract_S0016774614000390


How to cite this article:
D. Steinmetz, J. Winsemann, C. Brandes, B. Siemon, A. Ullmann, H. Wiederhold and U. Meyer Towards an improved
geological interpretation of airborne electromagnetic data: a case study from the Cuxhaven tunnel valley and its Neogene
host sediments (northwest Germany). Netherlands Journal of Geosciences, Available on CJO 2014 doi:10.1017/njg.2014.39
Request Permissions : Click here

Downloaded from http://journals.cambridge.org/NJG, IP address: 194.95.112.91 on 14 Jan 2015

Netherlands Journal of Geosciences Geologie en Mijnbouw page 1 of 27

doi:10.1017/njg.2014.39

Towards an improved geological interpretation of airborne


electromagnetic data: a case study from the Cuxhaven tunnel
valley and its Neogene host sediments (northwest Germany)

D. Steinmetz1 , J. Winsemann1, , C. Brandes1 , B. Siemon2 , A. Ullmann2,3 , H. Wiederhold3


& U. Meyer2
1

Institut fur
Geologie, Leibniz Universitat Hannover, Callinstr. 30, 30167 Hannover, Germany

Bundesanstalt fur
Geowissenschaften und Rohstoffe, Stilleweg 2, 30655 Hannover, Germany

Leibniz-Institut fur
Angewandte Geophysik, Stilleweg 2, 30655 Hannover, Germany

Corresponding author. Email: winsemann@geowi.uni-hannover.de

Abstract
Airborne electromagnetics (AEM) is an effective technique for geophysical investigations of the shallow subsurface and has successfully been
applied in various geological settings to analyse the depositional architecture of sedimentary systems for groundwater and environmental purposes.
However, interpretation of AEM data is often restricted to 1D inversion results imaged on resistivity maps and vertical resistivity sections. The
integration of geophysical data based on AEM surveys with geological data is often missing and this deficiency can lead to uncertainties in the
interpretation process. The aim of this study is to provide an improved methodology for the interpretation of AEM data and the construction
of more realistic 3D geological subsurface models. This is achieved by the development of an integrated workflow and 3D modelling approaches
based on combining different geophysical and geological data sets (frequency-domain helicopter-borne electromagnetic data (HFEM), time-domain
helicopter-borne electromagnetic data (HTEM), three 2D reflection seismic sections and 488 borehole logs). We used 1D inversion results gained from
both HFEM and HTEM surveys and applied a 3D resistivity gridding procedure based on geostatistical analyses and interpolation techniques to create
continuous 3D resistivity grids. Subsequently, geological interpretations have been performed by combining with, and validation against, borehole
and reflection seismic data. To verify the modelling results and to identify uncertainties of AEM inversions and interpretation, we compared the
apparent resistivity values of the constructed 3D geological subsurface models with those of AEM field measurements. Our methodology is applied
to a test site near Cuxhaven, northwest Germany, where Neogene sediments are incised by a Pleistocene tunnel valley. The Neogene succession is
subdivided by four unconformities and consists of fine-grained shelf to marginal marine deposits. At the end of the Miocene an incised valley was
formed and filled with Pliocene delta deposits, probably indicating a palaeo-course of the River Weser or Elbe. The Middle Pleistocene (Elsterian)
tunnel valley is up to 350 m deep, 0.82 km wide, and incised into the Neogene succession. The unconsolidated fill of the Late Miocene to
Pliocene incised valley probably formed a preferred pathway for the Pleistocene meltwater flows, favouring the incision. Based on the 3D AEM
resistivity model the tunnel-valley fills could be imaged in high detail. They consist of a complex sedimentary succession with alternating fine- and
coarse-grained Elsterian meltwater deposits, overlain by glaciolacustrine (Lauenburg Clay Complex) and marine Holsteinian interglacial deposits.
The applied approaches and results show a reliable methodology, especially for future investigations of similar geological settings. The 3D resistivity
models clearly allow a distinction to be made between different lithologies and enables the detection of major bounding surfaces and architectural
elements.
Keywords: 3D subsurface model, airborne electromagnetics, Cuxhaven, Neogene incised valley, Pleistocene tunnel valley, uncertainties, workflow

Introduction
Airborne electromagnetics (AEM) is an effective technique
to investigate the shallow subsurface. It has successfully
been applied in various geological settings to analyse the


C

Netherlands Journal of Geosciences Foundation 2015

depositional architecture (e.g. Jordan & Siemon, 2002; Huuse


et al., 2003; Sandersen & Jrgensen, 2003; Paine & Minty, 2005;
Jrgensen & Sandersen, 2006, 2009; Bosch et al., 2009; Steuer
et al., 2009; Pryet et al., 2011; Burschil et al., 2012a,b; Klimke
et al., 2013). AEM enables a fast geological overview mapping

Netherlands Journal of Geosciences Geologie en Mijnbouw

of subsurface structures and allows different lithologies and


pore water conditions to be distinguished (Siemon, 2005; de
Louw et al., 2011; Burschil et al., 2012a). Inversion procedures are carried out to create resistivity-depth models (e.g.
Siemon et al., 2009a,b), which are the basis for resistivitydepth sections and maps. Although this technique is a wellestablished method to improve geological interpretations of the
subsurface architecture, there is a substantial need for an efficient and reliable methodology to image the results in three
dimensions.
Progress in imaging the results in three dimensions was made
by combining 1D inversion models into 3D gridded data models
(e.g. Lane et al., 2000; Jrgensen et al., 2005; Bosch et al., 2009;
Palamara et al., 2010; Jrgensen et al., 2013). However, this integration gave limited consideration to uncertainties related to
the interpolation procedure (e.g. Pryet et al., 2011). To minimise these uncertainties different approaches were developed,
including an integrated geophysical and geological interpretation based on AEM surveys, reflection seismic sections, borehole data and logs. This provides the most reliable results and
leads to a minimisation of interpretational uncertainties (e.g.
Gabriel et al., 2003; Jrgensen et al., 2003a; BurVal Working
Group, 2009; Jrgensen & Sandersen, 2009; Hyer et al., 2011;
Jrgensen et al., 2013). However, little attention has been paid
to developing methodologies with an integrated interpretation
using different geological and geophysical data sets for the
shallow subsurface.
The aim of this paper is to provide a methodology to construct a 3D subsurface model with reduced interpretation uncertainties by integrating AEM, borehole and seismic data. The
method was applied and tested with data sets 6 km south from
Cuxhaven, northwest Germany. The study area comprises Neogene sediments that are incised by an Elsterian tunnel valley.
Previous studies focused on overview mapping of Neogene and
Palaeogene marker horizons and Pleistocene tunnel valleys from
2D reflection seismic sections and borehole logs (Gabriel et al.,
2003; BurVal Working Group, 2009; Rumpel et al., 2009). Larger
conductive structures within the valley fill were identified from
1D inversion results of frequency-domain helicopter-borne electromagnetic data (HFEM) and time-domain helicopter-borne
electromagnetic data (HTEM) inversion results (Rumpel et al.,
2009; Steuer et al., 2009). Using the previous results and interpretations as well as our new data sets, we provide a detailed
analysis of the seismic facies and sedimentary systems included
in a 3D geological subsurface model, resolving architectural elements in much greater detail. We generated 3D resistivity grids
based on the geostatistical analysis and interpolation of 1D AEM
inversion results. The 3D resistivity grids combine the advantage of volumetric computations with the visualisation of wide
resistivity ranges and allow the direct comparison and implementation of additional data sets such as borehole and seismic
data to improve the geological interpretation of the shallow
subsurface.

Geological setting and previous research


The study area is 7.5 km by 7.5 km and is located in northwest
Germany, between Cuxhaven in the north and Bremerhaven
in the south (Fig. 1A). The study area belongs to the Central
European Basin System that evolved from the Variscan foreland
basin in the Late Carboniferous (Betz et al., 1987).
In Permian times a wide continental rift system developed,
resulting in NS trending graben structures (Gast & Gundlach,
2006). During the Late Permian, repeated marine transgressions
flooded the subbasins and thick evaporite successions formed
(Pharaoh et al., 2010). Extensional tectonics during the Middle to Late Triassic led to the formation of NNESSW trending
graben structures, following the orientation of major basement
faults. The main extensional phase in the Late Triassic was accompanied by strong salt tectonics and rim-syncline development (Kockel, 2002; Grassmann et al., 2005; Maystrenko et al.,
2005a). During the Late Cretaceous to Early Palaeogene, the
area was tectonically reactivated, an event that is related to
the Alpine Orogeny (Maystrenko et al., 2005a) and was accompanied by local subsidence and ongoing salt tectonics (Baldschuhn et al., 1996, 2001; Kockel, 2002; Grassmann et al., 2005;
Maystrenko et al., 2005a; Rasmussen et al., 2010).

Palaeogene and Neogene marginal-marine deposits


Since the Late Oligocene, sedimentation in the North Sea Basin
has been dominated by a large clastic depositional system, fed
by the Baltic River System (Huuse & Clausen, 2001; Overeem
et al., 2001; Huuse, 2002; Kuster, 2005; Mller et al., 2009; Knox
et al., 2010; Anell et al., 2012; Rasmussen & Dybkjr, 2013;
Thole et al., 2014). Initially the Baltic River System, draining
the Fennoscandian Shield and the Baltic Platform, prograded
from the northeast, and then the sediment transportation direction rotated clockwise to southeast (Srensen et al., 1997;
Michelsen et al., 1998; Huuse et al., 2001).
Early Miocene deposits in the study area consist of finegrained glauconite-rich marine outer shelf deposits bounded
at the base by an unconformity (Gramann & Daniels, 1988;
Gramann, 1989; Overeem et al., 2001; Kuster, 2005). KAr apparent ages of these sediments range from 24.8 to 22.6 Ma
(Odin & Kreuzer, 1988), indicating that the basal unconformity
correlates with the marked climatic deterioration and eustatic
sea-level fall at the PalaeogeneNeogene transition (Huuse &
Clausen, 2001; Zachos et al., 2001; Miller et al., 2005; Rasmussen et al., 2008, 2010; Anell et al., 2012).
The lower boundary of the Middle to Late Miocene deposits
is the intra Middle Miocene unconformity. The unconformity
can be traced as a strong seismic reflector or as a prominent
downlap surface in most parts of the North Sea Basin (Cameron
et al., 1993; Michelsen et al., 1995; Huuse & Clausen, 2001;
Rasmussen, 2004; Mller et al., 2009; Anell et al., 2012) and
coincides with a significant depositional hiatus in the southern,

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 1. (A) Location of the study area and maximum extent of the Pleistocene ice sheets (modified after Jaritz, 1987; Baldschuhn et al., 1996, 2001;
Scheck-Wenderoth & Lamarche, 2005; Ehlers et al., 2011). (B) Hill-shaded relief model of the study area, showing the outline of the Hohe Lieth ridge.
(C) Close-up view of the study area showing the location of Pleistocene tunnel valleys (light yellow) and the location of boreholes (coloured dots).
Boreholes used in the seismic and cross-sections are indicated by larger dots; seismic sections S1 (Fig. 4), S2 and S3 (Fig. 5) are displayed as black lines.
Cross-section AB is visualised in Fig. 6; CD is visualised in Fig. 7. (D) HFEM surveys are displayed as grey lines; HTEM surveys are displayed as dark grey
dots.

central and northern North Sea Basin (Rundberg & Smalley,


1989; Huuse & Clausen, 2001; Stoker et al., 2005a,b; Eidvin &
Rundberg, 2007; Kothe, 2007; Anell et al., 2012).
The Middle Miocene hiatus is characterised by sediment starvation and/or condensation, and might have resulted from

a relative sea-level rise, either eustatic or in combination


with tectonic subsidence (Gramann & Kockel, 1988; Cameron
et al., 1993; Anell et al., 2012). In the study area the age
of the Middle Miocene unconformity is dated to 13.214.8 Ma
(Kothe et al., 2008). The overlying fine-grained, glauconite-rich

Netherlands Journal of Geosciences Geologie en Mijnbouw

Middle Miocene shelf deposits are characterised by an overall


fining-upward trend (Gramann & Daniels, 1988; Gramann, 1989;
Overeem et al., 2001; Kuster, 2005).
The Middle Miocene climatic optimum corresponds to a sealevel highstand, which was followed by a climatic cooling and
an associated sea-level fall during the late Middle Miocene (Haq
et al., 1987; Jurgens,

1996; Zachos et al., 2001; Kuster, 2005;


Miller et al., 2005). In the study area, Late Miocene deposits
consist of shelf and storm-dominated shoreface deposits with
an overall coarsening-upward trend (Gramann & Kockel, 1988;
Gramann, 1988, 1989; Kuster, 2005) and KAr apparent ages
ranging between 9.5 and 11 Ma (Odin & Kreuzer, 1988).
The Pliocene was characterised by several cycles of transgression and regression, resulting in the formation of unconformities, which can be traced in most areas of the North Sea
Basin (Mangerud et al., 1996; Konradi, 2005; Kuhlmann & Wong,
2008; Thole et al., 2014).

Pleistocene deposits
The Early Pleistocene was characterised by marine to fluviodeltaic sedimentation in the study area (Gibbard, 1988;
Kuhlmann et al., 2004; Streif, 2004). Subsequently, the Middle
Pleistocene (MIS 12 to MIS 6) Elsterian and Saalian ice sheets
completely covered the study area (Fig. 1A); the Late Pleistocene Weichselian glaciation did not reach the study area (e.g.
Caspers et al., 1995; Streif, 2004; Ehlers et al., 2011; Roskosch
et al., 2014).
During the Elsterian glaciation, deep tunnel valleys were
subglacially incised into the Neogene and Pleistocene deposits
of northern Central Europe (Huuse & Lykke-Andersen, 2000;
Lutz et al., 2009; Stackebrandt, 2009; Lang et al., 2012; van der
Vegt et al., 2012; Janszen et al., 2012, 2013). In this period, the
0.82 km wide and up to 350 m deep Cuxhaven tunnel valley
formed (Kuster & Meyer, 1979; Ortlam, 2001; Wiederhold et al.,
2005a; Rumpel et al., 2009). The valley is filled with Elsterian
meltwater deposits and till (Gabriel, 2006; Rumpel et al., 2009),
overlain by the Late Elsterian glaciolacustrine Lauenburg Clay
Complex and marine Holsteinian interglacial deposits, which
consist of interbedded sand and clay (Kuster & Meyer, 1979,
1995; Linke, 1993; Knudsen, 1988, 1993a,b; Muller

& Hofle,
1994; Litt et al., 2007). 2D resistivity sections based on AEM
surveys indicate that depositional units of the valley fill can
vary in thickness over a short distance (Siemon et al., 2002;
Gabriel et al., 2003; Siemon, 2005; Rumpel et al., 2009; Steuer
et al., 2009; BurVal Working Group, 2009). The Elsterian and
interglacial Holsteinian deposits of the tunnel-valley fill and
the adjacent Neogene host sediments are overlain by Saalian
sandy meltwater deposits and till (Kuster & Meyer, 1979; Ehlers,
2011). The morphology of the study area is characterised by an
up to 30 m high Saalian terminal moraine complex, the Hohe
Lieth ridge (Fig. 1B; Ehlers et al., 1984). Eemian tidal salt marsh
deposits together with Holocene fluvial deposits, peats and soils

characterise the lowland on both sides of the Hohe Lieth ridge


(Fig. 1B; Hofle et al., 1985; Binot & Wonik, 2005; Panteleit &
Hammerich, 2005; Siemon, 2005).

Database
The database includes 488 borehole logs, three 2D reflection
seismic sections and two AEM surveys, comprising both HFEM
and HTEM data (Fig. 1C and D). Most of the data sets were acquired between 2000 and 2005 by the BurVal project (Blindow &
Balke, 2005; Wiederhold et al., 2005a,b; BurVal Working Group,
2009; Tezkan et al., 2009).

Borehole data and geological depth maps


R
ParadigmTM GOCAD
software (Paradigm, 2011) was used to
construct an initial 3D geological subsurface model from a
Digital Elevation Model (DEM) with a resolution of 50 m and
published depth maps of the Cenozoic succession (Baldschuhn
et al., 1996; Rumpel et al., 2009; Hese, 2012). Lithology logs
of 488 boreholes were used to analyse the subsurface architecture of the Cenozoic deposits and to define major geological
R
(Fugro Conunits. The commercial software package GeODin
sult GmbH, 2012) was used for the data management.
Borehole location data were corrected for georeferencing errors. The interpretation of the tunnel-valley fill is based on
five borehole logs penetrating all five lithologic facies units
(Fig. 1C). A resistivity log was only available for borehole Hl9
Wanhoeden located in the centre of our test site, penetrating
the Cuxhaven tunnel valley (Fig. 1C).

Seismic sections
The seismic surveys were acquired by the Leibniz Institute for
Applied Geophysics Hannover (LIAG, formerly the GGA Institute) in 2002 and 2005. The three 2D reflection seismic sections
were used to map the large-scale subsurface architecture of the
study area (black lines in Fig. 1C). The 2D reflection seismic
sections include a 6 km long WNWESE oriented seismic line S1
and a 2.4 km long WE trending seismic line S2 that is located
1 km further to the south (Fig. 1C). Seismic line S2 intersects
with the 0.4 km long, NS trending seismic line S3. The survey design and the processing of the seismic sections, lines S1
(Wanhoeden), S2 (Midlum 3) and S3 (Midlum 5), are described
by Gabriel et al. (2003), Wiederhold et al. (2005b) and Rumpel
et al. (2006a,b; 2009).
R
Landmark
Data processing was carried out using ProMax
software. Processing followed the workflow for vibroseis data
described by Yilmaz (2001). The seismic vibrator operated with
a sweep ranging from 50 to 200 Hz. The survey design led to
a maximum target depth of 1600 m with a common-midpoint
spacing of 5 m for seismic line S1 and 2.5 m for seismic lines

Netherlands Journal of Geosciences Geologie en Mijnbouw

S2 and S3. Assuming a velocity of 1600 m/s and a maximum


frequency of about 150 Hz, wavelengths of about 10 m are expected and thus the minimum vertical resolution should be in
the range of a quarter of a wavelength but at least about 4 m for
the shallow subsurface. Increasing velocity and decreasing frequency with depth leads to a decrease in resolution with depth
(wavelength of about 22 m in 1000 m depth). For migration as
well as for depth conversion a simple smooth velocity function
is used with a start velocity of 1500 m/s increasing to about
2200 m/s at 1000 m depth.

Acquisition, processing and visualisation of AEM


data
Frequency-domain helicopter-borne electromagnetics The HFEM
survey covering the entire study area (Fig. 1D) was conducted
by the Federal Institute for Geosciences and Natural Resources
(BGR) in 2000. The survey grid consisted of parallel WNWESE
flight lines with an average NNESSW spacing of 250 m, connected by tie lines perpendicular to the flight lines with a WNW
ESE spacing of 1000 m (grey lines in Fig. 1D). The distance
between consecutive values was 34 m, assuming an average
flight velocity of about 140 km/h during the survey (Siemon
et al., 2004). The HFEM system comprised a five-frequency device (Siemon et al., 2002). The transmitter/receiver coil configuration was horizontal co-planar for all frequencies. The
transmitter coils operated at frequencies of 0.4 kHz, 1.8 kHz,
8.6 kHz, 41.3 kHz and 192.6 kHz (Siemon et al., 2004). The
maximum penetration depth was about 150 m, depending on
the subsurface resistivity distribution. The sampling rate was
0.1 s and the signal was split into its in-phase I and out-ofphase Q components relative to the transmitter signal. Because
of the relatively small system footprint (about 100150 m), i.e.
the lateral extent of the main inductive response beneath the
system, the data set is characterised by a rather high spatial
resolution, which decreases with depth (Siemon et al., 2004).
The inversion of HFEM data to resistivity and depth values followed the workflow developed in Sengpiel & Siemon (2000) and
Siemon (2001). It depended on an initially unknown subsurface
resistivity distribution. Initially the resistivity was calculated
based on a half-space model (Fraser, 1978). If the resistivity
varies with depth, the uniform half-space model will yield different apparent resistivity and apparent distance values at
each HFEM frequency (Siemon, 2001). The centroid depth is
a measure for the penetration of the electromagnetic fields
and represents the centre of the half-space. This depth value
depends on the individual HFEM frequencies and on the resistivity distribution in the subsurface: the higher the ratio of
resistivity and frequency, the greater the centroid depth. The
apparent resistivity and centroid depth data pairs were used
to determine a set of sounding curves at each data point. The
apparent resistivity vs centroid depth-sounding curves are a
smooth approximation of the vertical resistivity distribution

and were also used to define an individual six-layer starting


model at each data point for an iterative MarquardtLevenberg
inversion. The model parameters were modified until a satisfactory fit between the survey data and the calculated field data
from the inversion model was achieved (Sengpiel & Siemon,
2000; Siemon et al., 2009a,b). Based on the low number of
input parameters available (two per frequency), which can be
resolved by 1D inversion, the number of individual model layers is limited. The 1D HFEM inversion results were extracted as
a set of data points (reduced to approximately one sounding
or 1D model every 7 m along the flight lines). The apparent
resistivities are displayed separately at each frequency.
Time-domain helicopter-borne electromagnetics An HTEM survey
was conducted by the University of Aarhus (BurVal Working
Group, 2009; Rumpel et al., 2009) on behalf of the Leibniz
Institute for Applied Geophysics (LIAG, formerly the GGA Institute). The HTEM soundings are restricted to an area of about
3.96 km by 2.8 km (grey dots in Fig. 1D). The HTEM system operated with a transmitter loop on a six-sided frame (Srensen
& Auken, 2004). In the Cuxhaven survey, the distance between
consecutive soundings was about 75 m, assuming an average
flight velocity of 18 km/h during the survey (Fig. 1D).
The system operated with low and high transmitter moments. A low transmitter moment of approximately 9000 Am2
was generated by a current of 40 A in one measurement cycle.
The received voltage data were recorded in the time intervals
between 17 and 1400 s. A current of 4050 A in four loop
turns generated a high moment of approximately 47,000 Am2 .
The voltage data of high moment measurements were recorded
in the time interval of 1503000 s. Acquisition for both the
low and high transmitting moments were carried out in cycles for the four data sets (320 stacks per data set for low
and 192 stacks per data set for high moment measurement).
Subsequently, the arithmetic mean values of the low and high
moments were averaged for each of the data sets to generate
one data set. This merged data set was interpreted as one geophysical model. The resolution of the uppermost part of the
subsurface is limited by the recording lag time of the HTEM
system, which does not start until about 17 s, and strongly
depends on the near surface conductivity and thus on lithology
and the pore water content (Steuer et al., 2009). This results
in near-surface layers being commonly merged into one layer
in the model.
The maximum penetration depth is about 250 m and depends
on the subsurface resistivity. Because the method cannot resolve thin depositional units (<5 m), individual depositional
units may be merged into thicker model layers. With increasing
depth the resolution decreases, thus restricting the vertical resolution to approximately 1520 m in the shallow subsurface and
decreasing to 2050 m at 100 m depth. The horizontal resolution capability also decreases with depth. At about 25 m depth,
the diameter of the footprint from which data are obtained is

Netherlands Journal of Geosciences Geologie en Mijnbouw

R
used ParadigmTM GOCAD
software for 3D subsurface modelling
(Paradigm, 2011).

Construction of an initial 3D geological subsurface


model based on the interpretation of borehole and
seismic data

Fig. 2. Workflow.

about 75100 m and it exceeds 300400 m at 100 m depth (West


& Macnae, 1991; Jrgensen et al., 2005, 2013). Small-scale spatial variations in geology are therefore less well resolved at
deeper levels than at shallower levels (Newman et al., 1986).
Aarhus Workbench software was used to process the HTEM
data (Steuer, 2008; Steuer et al., 2009). This software integrates all steps in the processing workflow from management
of the raw data to the final visualisation of the inversion results. The software provides different filtering and averaging
tools, including the correction of GPS signal, tilt and altitude
values. HTEM data were inverted with a five-layer model using a spatially constrained inversion (SCI) option of em1dinv
(Steuer, 2008; Viezzoli et al., 2008; Steuer et al., 2009; HGG,
2011). SCI takes many adjacent data sets into account, which
are connected by distant dependent lateral constraints to impose continuity in areas with sparser data coverage. The lateral
constraints were defined by the expected geological continuity of each layer. The inversion results strongly depended on
the starting model used and the SCI settings, especially the
strength of constraints.
The 1D HTEM inversion results were extracted as a set of
models, which were characterised by a limited spatial resolution
due to the large distance between soundings and the relatively
large lateral extent of the main inductive response beneath the
system.

Methodology of combined geological and


geophysical analysis
To improve the interpretation of the subsurface architecture
we developed a workflow in which the interpretation results of borehole lithology logs, 2D reflection seismic sections
and 3D resistivity grids based on 1D AEM inversion results
were integrated (Figs 2 and 3). In order to achieve this we

In the first step a basic subsurface model of the Cuxhaven


tunnel valley and its Neogene host sediments was constructed
by integrating borehole lithology logs and 2D reflection seismic sections (Tables 1 and 2; Figs 3A and 47). For the analysis of seismic sections we used the scheme of Mitchum et al.
(1977). Each seismic unit is defined by the external geometry,
the internal reflector configuration and seismic facies parameters, such as amplitude, continuity and density of reflectors
(Tables 1 and 2; Figs 4 and 5). The 2D seismic interpretation
results combined with borehole lithology analysis defined 15
stratigraphic marker horizons, which represent the tops of each
depositional unit (Tables 1 and 2).
In GOCAD we applied a discrete modelling approach that
creates triangulated surfaces from points, lines, and open and
closed curves (Mallet, 2002). With the Discrete Smooth Interpolation (DSI) algorithm the roughness of the triangulated surfaces was minimised (Mallet, 2002). The 3D subsurface model
consists of a series of triangulated surfaces representing the
bounding surfaces of the depositional units (cf. Tables 1 and 2;
Fig. 3).

Construction of 3D resistivity grids based on AEM


data
The best technique to transform subsurface resistivity data from
a set of 1D vertical inversion models into a 3D model is 3D
interpolation. The applied 3D interpolation algorithm requires
discrete data in all directions, discarding the layered approach
used in the inversion, and leads to a smoothing effect between
previously defined layer boundaries of 1D AEM inversion.
In the first step, the 1D AEM inversion models were imported into the GOCAD modelling software (Fig. 3B). Subsequently, each AEM data set was transformed in a regular spaced
voxel grid with rectangular hexahedral cells.
The large differences in AEM data distribution and penetration depths required the construction of two independent grid
models with different resolutions. For the data set of 1D HFEM
inversion results we choose a grid with horizontal and vertical
cell sizes of 10 m and 1 m, respectively, which provides a good
compromise between data resolution and model handling. For
the data set of 1D HTEM inversion results an enlarged grid with
horizontal and vertical cell sizes of 20 m and 2.5 m, respectively, provides the best results.
The arithmetic mean was used to transform and estimate
the resistivity value of a cell based on each 1D AEM inversion
model. The estimation of resistivity at unsampled locations to

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 3. Workflow of the construction of combined geological/geophysical subsurface models, illustrated by HFEM data. (A) Construction of the 3D geological subsurface model based on borehole and seismic data. (B) Construction of a continuous 3D resistivity voxel grid based on 1D HFEM inversion
results. The resistivity data were integrated into a regular structured grid, analysed by means of geostatistical methods and subsequently interpolated.
(C) The selection of specific resistivity ranges provides a first estimate of the large-scale depositional architecture. Shown is the clay distribution in the study
area, indicated by HFEM resistivity values between 3 and 25 m. (D) Adjustment of the 3D geological subsurface model by integrating information of the
3D resistivity grid.

get a continuous 3D resistivity grid model of the subsurface


was achieved by the ordinary kriging method (Krige, 1951). A
1D vertical analysis combined with a 2D horizontal analysis of
the 1D AEM inversion models was carried out since sedimentary
successions tend to be more variable in the vertical than in the
horizontal direction and this often results in zonal anisotropy.
Kriging uses semivariogram models to infer the weighting given
to each data point and therefore takes both distance and direction into account. Variography was performed in different
directions (azimuths of 0, 45, 90 and 135) with a tolerance
of 22.5 and an adjusted bandwidth to survey the resistivity
isotropy in each resistivity model (Fig. 8). For each AEM data
set, best results in block weighting were obtained by using an
exponential function. The analysis of HFEM data results in a
vertical range of 25 m and a horizontal major principal axis

with an angle of 19 and a range of 1000 m, and a perpendicular minor principal axes with a range of 780 m (Fig. 8). The
analysis of HTEM data results in a vertical range of 50 m and
a horizontal major principal axes with an angle of 130 and a
range of 1500 m, and a perpendicular minor principal axes with
a range of 1300 m (Fig. 8). The 3D interpolation results for each
voxel over the whole study area in an estimated value for the
resistivity.
Uncertainties of 3D resistivity grid modelling
1) Data processing of vintage 1D AEM inversion models did
not focus on the elimination of anthropogenic effects (such
as the airport Cuxhaven/Nordholz in the northwest of the
study area). Hence, the processed AEM databases contained

Netherlands Journal of Geosciences Geologie en Mijnbouw

Table 1. Seismic and sedimentary facies of the Oligocene and Neogene marine to marginal marine deposits

Pliocene

Late

Mess.

Sedimentary
facies

Interpretation

Seismic
unit

Thickness
(m)

Hummocky, highamplitude reflectors


and
sub-parallel, partly
hummocky and
U-shaped, lowamplitude and
transparent reflectors

Fine sand and


silt

Estuarine deposits
(Gramann, 1989;
Kuster, 2005) of an
incised valley

U4

30120

99

97

100

Sub-parallel, partly
hummocky, lowamplitude reflectors

Fine sand

1540

88

90

100

3085

73

75

40

Continuous, subparallel, partly


hummocky, highamplitude reflectors

B + S AEM Lith.

Open shelf to stormdominated deposits


of the upper and
lower shoreface
(Gramann, 1989;
Rasmussen et al.,
2010; Rasmussen &
Dybkjr, 2013)

U3

Glauconitic clay
and silt

Transgressive shelf
deposits
(Gramann, 1989;
Kuster, 2005; Kthe,
2008; Rasmussen
et al., 2010)

U2

1070

Glauconitic clay,
silt and fine sand

Marine outer shelf


deposits
(Gramann, 1989;
Overeem et al.,
2001; Kuster, 2005)

U1

1085

U0

~220

Clay and silt


coarsening
upwards into fine
sand

Middle

Miocene

Tort.

(m)

Seismic pattern

Continuous, parallel,
partly hummocky,
Serra.
high-amplitude
reflectors

Early

Lang.

Continuous, subhorizontal, parallel,


Burg. partly hummocky,
high-amplitude
Aquit.
reflectors

Oligocene

Continuous, subhorizontal, highamplitude reflectors


and discontinuous,
partly inclined, lowamplitude reflectors,
partly transparent

Marine deposits
Clay, silt and
(Gramann, 1989; Kuster,
fine sand
2005)

Median resistivity values are extracted from the HFEM grid model and are related to the 3D geological subsurface model based on borehole and
seismic data (B + S), the adjusted 3D geological subsurface model derived from the 3D AEM resistivity grids (AEM) and the adjusted 3D geological
subsurface model with manually adjusted resistivity values based on lithology log information (Lith.).

erroneous data from anthropogenic noise (Siemon et al.,


2011).
2) 1D AEM inversion models are always simplified realisations
of the subsurface resistivity distribution, particularly if only
models with few layers are used to explain the AEM data.
As AEM resolution capability decreases with depth, the

layer thicknesses generally increase with depth, which corresponds to the probability that several thin layers of various
depositional units may be merged to a thicker resistivity
layer.
3) Uncertainties may originate from the interpolation method
(Pryet et al., 2011). We focused on uncertainties caused by

Netherlands Journal of Geosciences Geologie en Mijnbouw

Table 2. Seismic and sedimentary facies of the Pleistocene deposits

Seismic pattern

Thickness is below the seismic resolution

Seismic
unit

Salt marsh deposits


Fine sand and
(Hfle et al., 1985;
peat, contain clay
Knudsen, 1988; Gabriel,
and silt
2006)

Thickness
(m)

(m)
B + S AEM Lith.

up to 9

64

128

64

60

138
480

140
500

Hummocky, low- to
high-amplitude
reflectors passing
laterally into
continuous,
horizontal-parallel or
parallel-inclined
reflectors

Fine to
coarse sand
and
diamicton

Diamicton,
glaciofluvial and
terminal moraine
deposits
(Ehlers, 2011)

U5.8
sat.
unsat.

1050

Holsteinian

Eem.

Interpretation

Saalian
Drenthe

Late

Holocene
Wei.

Sedimentary
facies

Discontinuous,
hummocky to
inclined-parallel,
high- to mediumamplitude reflectors

Clay, silt
and fine
sand with
shells

Marine to marginal
marine deposits
(Kuster & Meyer, 1979;
Knudsen, 1988)

U5.7
U5.7*

645

24
69

32

30

1045

16

14
52

10
10

Elsterian

Pleistocene
Middle

Horizontal-parallel
reflectors passing
laterally into
discontinuous,
hummocky high- to
medium-amplitude
reflectors

Glaciolacustrine deposits
Clay and silt
of the Upper Lauenburg U5.6
with some fine
Clay Complex
U5.6*
sand
(Kuster & Meyer, 1979)

Fine to medium
sand, silty

Glaciolacustrine
deposits of the
Lower Lauenburg
Clay Complex
(Kuster & Meyer,
1979)

U5.5
U5.5*

2035

29
100

25
91

25
70

Fine to medium
sand, delta
deposit

Glaciolacustrine
deposits of the
Lower Lauenburg
Clay Complex
(Kuster & Meyer,
1979; Ortlam, 2001)

U5.4

up to 50

33

34

150

Discontinuous,
parallel, hummocky,
medium- to lowamplitude reflectors,
partly transparent

Clay and silt

Glaciolacustrine
deposits of the
Lower Lauenburg
Clay Complex
(Kuster & Meyer,
1979)

U5.3

up to 55

31

27

15

Parallel-inclined,
hummocky,
subhorizontal, low- to
high-amplitude
reflectors

Fine sand fining


upwards to silty
clay

Fine-grained
glaciofluvial
deposits
(Kuster & Meyer,
1979)

U5.2
U5.2*

2565

38
66

35
58

35
35

Discontinuous, partly
inclined and
hummocky, high- to
medium-amplitude
reflectors

Coarse sand and


gravel fining
upwards into
medium sand

Coarse-grained
glaciofluvial
deposits, basal
tunnel-valley fill
(Kuster & Meyer,
1979)

U5.1

70160

41

33

120

Fine to medium
sand, silt,
pebbles

Proglacial meltwater
deposits

up to 10

67

66

70

Continuous to
discontinuous,
hummocky, parallel,
medium- to highamplitude reflectors

Discontinuous,
hummocky, low- to
high-amplitude
reflectors

Thickness is below the seismic resolution

Median resistivity values are extracted from the HFEM grid model and are related to the 3D geological subsurface model based on
borehole and seismic data (B + S), the adjusted 3D geological subsurface model derived from the 3D AEM resistivity grids (AEM)
and the adjusted 3D geological subsurface model with manually adjusted resistivity values based on lithology log information
(Lith.). Seismic subunits U5.1U5.8 refer to the Cuxhaven tunnel-valley fill. Values marked with refer to the small-scale tunnel
valley east of the Cuxhaven tunnel valley.

Netherlands Journal of Geosciences Geologie en Mijnbouw

10
Fig. 4. 2D reflection seismic section S1 combined with airborne electromagnetic data. (A) 2D reflection seismic section S1. (B) Interpreted seismic section S1 with borehole logs. Seismic units are described in Tables 1
and 2. (C) 2D reflection seismic section S1 combined with resistivity data, extracted from the 3D HFEM resistivity grid. The dashed line indicates the groundwater table. (D) 2D reflection seismic section S1 combined
with resistivity data, extracted from the 3D HTEM resistivity grid. For location see Fig. 1C and D.

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 5. 2D reflection seismic sections S2 and S3 combined with airborne electromagnetic data. (A) 2D reflection seismic sections S2 and S3. (B) Interpreted
seismic sections S2 and S3 with borehole logs. Seismic units are described in Tables 1 and 2. (C) 2D reflection seismic sections S2 and S3 combined with
resistivity data, extracted from the 3D HFEM resistivity grid. The dashed line indicates the groundwater table. (D) 2D reflection seismic sections S2 and S3
combined with resistivity data, extracted from the 3D HTEM resistivity grid. For location see Fig. 1C and D.

11

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 6. 2D cross-section of the study area (AB in Fig. 1C and D), showing major bounding surfaces and resistivity values based on borehole, seismic and
AEM data. The dashed line indicates the groundwater table. (A) 2D cross-section extracted from the 3D geological model based on borehole and seismic data.
Only the Pleistocene deposits are shown in colour. (B) 2D cross-section extracted from the adjusted 3D geological model and the corresponding resistivities
derived from the 3D HFEM voxel grid. (C) 2D cross-section extracted from the adjusted 3D geological model and the corresponding resistivities derived from
the 3D HTEM voxel grid. (D) 2D cross-section extracted from the adjusted 3D geological model. Only the Pleistocene deposits are shown in colour.

12

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 7. 2D cross-section of the study area (CD in Fig. 1C and D), showing major bounding surfaces and resistivity values based on borehole, seismic and
AEM data. The dashed line indicates the groundwater table. (A) 2D cross-section extracted from the 3D geological model based on borehole and seismic data.
Only the Pleistocene deposits are shown in colour. (B) 2D cross-section extracted from the adjusted 3D geological model and the corresponding resistivities
derived from the 3D HFEM voxel grid. (C) 2D cross-section extracted from the adjusted 3D geological model and the corresponding resistivities derived from
the 3D HTEM voxel grid. (D) 2D cross-section extracted from the adjusted 3D geological model. Only the Pleistocene deposits are shown in colour.

13

Netherlands Journal of Geosciences Geologie en Mijnbouw

the kriging interpolation method of the 1D AEM inversion


results. A geostatistical uncertainty estimate is provided by
the kriging variance 2 KRI (x, y, z), which depends on the
spatial variability of the parameter and the distance to individual data points. This leads to the construction of a 3D grid
that contains both resistivity values and their uncertainties
(Pryet et al., 2011). Once the model is built, the analysis of
uncertainties between data points (i.e. between flight lines)
is expressed by the standard deviation KRI (Fig. 9). The
standard deviation can be used to evaluate the credibility of
the interpolated data and to eventually exclude soundings
or groups of soundings with high uncertainty and instead
emphasise high-quality soundings. A value close to zero indicates high probability and a value close to one indicates
low probability of resistivity values. This information can be
used to evaluate the relationship between borehole lithology
and 3D resistivity data and hence to quantify uncertainties
in the depositional architecture.

Relationship between lithology and resistivity


Borehole lithology and AEM resistivity were linked with the aim
of using resistivity as a proxy for lithology (see: Bosch et al.,
2009; Burschil et al., 2012b). The resistivity log from borehole
Hl9 (Fig. 10) and sediment descriptions from 487 borehole logs
(Fig. 1C) were used to combine these parameters (Fig. 11).
The electrical resistivity of sediments is mainly controlled by
the presence of clay minerals, the degree of water saturation
and the pore water ion content. According to Archie (1942),
resistivity for clay-free sediments is inversely proportional to
the pore water ion content. If the specific pore water ion content is known throughout the different geological layers and
structures, estimates in lithology variations related to the clay
content and type can be obtained. Because saline groundwater
is absent in the study area, at least at depths resolved by HFEM
(Siemon, 2005), resistivity changes are related to changes in
lithology.
The lithologies were divided into seven grain-size classes
based on the borehole descriptions. In accordance with data
resolution, we determined the representative lithology in each
borehole and extracted the corresponding interpolated resistivity at this location from the 3D resistivity grids based on the two
different AEM data sets. We accept in this step that the lithology logs within the study area have varying resolutions ranging
from 1 cm to 1 m, depending on the drilling method and data
collection. This may lead to a discrepancy between the lithology
log and the vertical resolution of the 3D AEM resistivity grids,
resulting in insufficient resistivitygrain size relationships. To
reduce the uncertainty between the lithology logs and resistivity grid we analysed all logs and calculated mean and median
resistivity values for each class that allowed derivation of grain
size from resistivity (Fig. 11). Grouping the data into resistivity
classes and counting the number of occurrences of each lithol-

14

Fig. 8. Experimental vertical and horizontal semivariograms derived from


the 1D HFEM and HTEM inversion results of the study area. Variography
was performed in different directions (azimuths of 0, 45, 90 and 135)
with a tolerance of 22.5. The manually fitted exponential models are also
indicated.

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 9. 3D view of kriging uncertainty on log-transformed resistivity. Displayed is the kriging standard deviation KRI from the interpolation process of HFEM data. Low uncertainty traces (dark-blue) indicate the flight
lines while the greatest uncertainty is across lines (rather than along
lines).

Fig. 10. Seismic section (A), resistivity log (B) and lithology log (C) of borehole Hl9 Wanhoeden (after Besenecker, 1976). The blue curve shows the
measured resistivity log of the borehole, the green curve displays the resistivity log extracted from the 3D HFEM resistivity grid and the red curve displays the
projected resistivity log extracted from the 3D HTEM resistivity grid.

ogy resulted in the proportion of occurrences of each grain-size


class per resistivity group (Fig. 11). As high resistivity values
in the study area often represent dry sediments, we focused on
resistivity values up to 250 m.

Integration of the 3D resistivity grids with the 3D


geological subsurface model
To test the match between the 3D resistivity grid models (based
on 1D inversion results of AEM data) and the 3D geological
subsurface model (based on borehole and seismic data), we integrated both into GOCAD (Fig. 3C). To verify the validity and
accuracy of data and their integration we followed a mutual
comparison between seismic reflector pattern and the corresponding resistivities obtained from the 3D AEM resistivity grids
in GOCAD (Figs 4C and D Fig. 5C and D) and related grain-size

classes defined from borehole logs to resistivity values derived


from the 3D AEM resistivity grid models (Fig. 11).

Adjustment of the 3D geological subsurface model


The GOCAD modelling software has various possibilities to visualise properties such as resistivity values of the 3D grid models to facilitate use and interpretation of data. This includes
changes in the colour scale, the selection of given resistivity
ranges and voxel volumes. The selection of specific resistivity
ranges in the 3D resistivity grid model, in particular, provides a
first rough estimate of the three-dimensional geological architecture (Fig. 3D).
In this study, we propose a method for 3D geological modelling of AEM data in which the limitations are jointly considered. The relationship between lithology and resistivity and
its corresponding resistivity values was used to adjust the

15

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 11. Grain-size classes and related resistivity histograms extracted from the interpolated 3D HFEM and HTEM resistivity grid. A. Grain-size classes
and related resistivity extracted from the interpolated 3D HFEM and HTEM resistivity grid. Shown are mean values, standard deviation, median values
and number of counts for common resistivity (logarithmic values). B. Histogram showing resistivity classes of HFEM and HTEM data for each grainsize class as stacked bars (linear scale). C. Histogram showing resistivity classes of HFEM and HTEM data for each grain-size class as stacked bars
(logarithmic scale). The absence of resistivity values lower than 1 log10 m indicates that the influence of anthropogenic noise and saltwater can be
excluded.

bounding surfaces of the geological framework model. After


adjusting the geologic framework the updated 3D geological
subsurface models were combined with the 3D HFEM voxel grid.
Voxel grids can hold an unlimited number of attributes and
different parameters can be added to the grid structure as attributes such as lithology, facies or model uncertainty. The
high resolution of the regular voxel model allows the subsurface structure to be maintained. The result is a more reliable reconstruction of the shallow subsurface architecture (Figs 47).
Nevertheless, internal lithology variations as identified on seis-

16

mic sections and the heterogeneity of sediments represented by


overlapping resistivity ranges often leads to ambiguous lithological interpretation.

Verification of the 3D geological subsurface models


by means of HFEM forward modelling
We compared the apparent resistivity values at each frequency
of the HFEM survey (Fig. 12A) with the apparent resistivity
values derived from the initial 3D geological subsurface model

Netherlands Journal of Geosciences Geologie en Mijnbouw

Fig. 12. Apparent resistivity images at different frequencies, corresponding to centroid depths, which increase from left to right. A. Apparent resistivity images of measured HFEM data. B. Apparent resistivity images extracted from the 3D geological subsurface model based on borehole and seismic
data. C. Apparent resistivity images extracted from the adjusted 3D geological subsurface model derived from AEM data. D. Apparent resistivity images
extracted from the adjusted 3D geological subsurface model derived from AEM data with manually adjusted resistivity values based on lithology log
information.

17

Netherlands Journal of Geosciences Geologie en Mijnbouw

based on borehole and seismic data (Fig. 12B) and the adjusted
3D geological subsurface model derived from the 3D AEM resistivity grids (Fig. 12C). This allowed identification of differences
and uncertainties in each data set.
At each HFEM data point the thickness and median resistivity value of each geological unit derived from the different
3D geological subsurface models (Tables 1 and 2) were used to
create a 1D resistivity model. For all 1D models, synthetic HFEM
data were derived and transformed to apparent resistivity values at each HFEM frequency (Siemon, 2001). To verify the HFEM
data, apparent resistivities were used to compare the measured
and modelled HFEM data for two reasons: (1) the apparent resistivities are almost always independent of altitude variations
of the HFEM system (Siemon et al., 2009a) and (2) they represent an approximation to the true resistivity distribution in
the subsurface. The corresponding depth levels, however, vary
as the centroid depth values depend on the penetration depth
of the electromagnetic fields.

Results
The depositional architecture of the Cuxhaven
tunnel valley and its Neogene host sediments
defined by seismic and borehole data analysis
Neogene marine and marginal marine deposits The Neogene succession is 360 m thick and unconformably overlies open marine
to paralic Oligocene deposits (Kuster, 2005). On the basis of previous investigations of the Neogene sedimentary successions
(Gramann & Daniels, 1988; Odin & Kreuzer, 1988; Gramann,
1988, 1989; Overeem et al., 2001; Kuster, 2005; Kothe et al.,
2008; Rumpel et al., 2009) four seismic units were mapped
within the Neogene deposits, each bounded at the base by an
unconformity. These seismic units can be correlated to seismic
units in the North Sea Basin (Michelsen et al., 1998; Mller
et al., 2009; Anell et al., 2012). The main characteristics of
seismic units, including seismic facies, sedimentary facies and
mean resistivity values are summarised in Table 1. The overall thickening observed in the westward dip direction of the
seismic units is interpreted as an effect of the salt rim syncline subsidence creating accommodation space (cf. Maystrenko
et al., 2005a,b; Grassmann et al., 2005; Brandes et al., 2012).
The deposits consist of fine-grained shelf to marginal marine
sediments.
At the end of the Miocene an incised valley formed that
was subsequently filled with Pliocene delta deposits, probably
indicating a palaeo-course of the River Weser or Elbe (Figs 4,
5 and 6). This is confirmed by the sub-parallel pattern of the
lower valley fill onlapping reflector terminations onto the truncation surface observed in seismic lines S1 and S2. This is interpreted as transgressive backstepping (seismic lines S1 and S2;
Table 1; Figs 4A and B, and 5A and B; Dalrymple et al., 1992).

18

The upper part of the incised-valley fill is characterised in the


seismic by mound- and small-scale U-shaped elements, which
are interpreted as prograding delta lobes (seismic lines S1 and
S2; Figs 4A and B, and 5A and B).
Pleistocene and Holocene deposits Pleistocene deposits unconformably overly the marginal-marine Neogene sediments and
are separated at the base by an erosional surface characterised
by two steep-walled tunnel valleys passing laterally into subhorizontal surfaces (Fig. 4A and B). The large tunnel valley is
up to 350 m deep and 2 km wide (seismic line S1 and S2,
Figs 4A and B, and 5A and B) and has previously been described by Kuster & Meyer (1979), Siemon et al. (2002, 2004),
Gabriel et al. (2003), Siemon (2005), Wiederhold et al. (2005a),
Gabriel (2006), Rumpel et al. (2009) and BurVal Working Group
(2009). The smaller tunnel valley in the east is up to 200 m
deep and 1 km wide (seismic line S1, Fig. 4A and B). In total, eight seismic subunits were mapped within these tunnel
valleys and the marginal areas (U5.1U5.8; cf. Table 2; Figs 4B
and 5B). The main characteristics of seismic subunits, including
seismic facies, sedimentary facies and mean resistivity values,
are summarised in Table 2.
The dimension, geometry, internal reflector pattern and sedimentary fill of the troughs correspond well to Pleistocene
tunnel-valley systems described from northern Germany (Ehlers
& Linke, 1989; Stackebrandt, 2009; Lang et al., 2012; Janszen
et al., 2013), Denmark (Jrgensen & Sandersen, 2009), the
Netherlands (Kluiving et al., 2003) and the North Sea Basin
(Wingfield, 1990; Huuse & Lykke-Andersen, 2000; Praeg, 2003;
Lutz et al., 2009; Kehew et al., 2012; Moreau et al., 2012). The
basal fill consists of Elsterian meltwater deposits overlain by
glaciolacustrine deposits of the Lauenburg Clay Complex. These
deposits are unconformably overlain by marine Holsteinian interglacial deposits. Saalian meltwater deposits and till cover
the entire study area and unconformably overly the Neogene,
Elsterian and Holsteinian deposits (Table 2; Figs 4A and B, and
5A and B).

Integration of the resistivity grid model and the


geological subsurface model
Comparison of borehole resistivity logs with model resistivity data
The borehole Hl9 Wanhoeden penetrates the western marginal
tunnel-valley fill (cf. Figs 1C and 4B). The lowermost fill consists
of Elsterian glaciolacustrine fine- to medium-grained sand and
silt alternations of the Lower Lauenburg Clay Complex (U5.2
5.4). This succession is overlain by the Upper Lauenburg Clay
Complex, which consists of clay, silt and fine-grained sand alternations (U5.55.6) and marine Holsteinian interglacial deposits (U5.7; Figs 4B and 10). The resistivity log of borehole Hl9
displays strong resistivity variations, which can be correlated
with the alternation of clay- and sand-rich beds (Fig. 10). The
3D resistivity grid model based on 1D HFEM inversion results

Netherlands Journal of Geosciences Geologie en Mijnbouw

displays a similar resistivity log as the one in the borehole and


thus a good lithological match. The upper bounding surface of
the Upper Lauenburg Clay Complex and the marine interglacial
Holsteinian clay can also be identified on the 3D resistivity
grid model (Fig. 10). The interpolated 3D resistivity grid model
based on 1D HTEM inversion results, however, reveals only one
conductor. While the Upper Lauenburg Clay Complex is nearly
fully imaged, the interglacial marine Holsteinian clay is not
well defined, which may be caused by the limited resolution of
the transient electromagnetic (TEM) method. The highly conductive sediments of the Upper Lauenburg Clay Complex lead to
a reduced penetration of the electromagnetic fields and hence
resistivities of the sediments below the Upper Lauenburg Clay
Complex are not detectable with the transmitter moments used
in this survey.
An important difference between resistivity values of the
measured resistivity log of borehole Hl9 and the extracted values from the 3D resistivity grid models based on 1D AEM inversion results is the amplitude of high resistivity values, which is
considerably lower in the 3D resistivity grid models (Fig. 10).
This difference can be explained by the applied AEM methods,
which are more sensitive to conductive sediments (Steuer et al.,
2009). In addition, the AEM system predominantly generates
horizontal currents, whereas electromagnetic borehole systems
use vertical currents for measuring the log resistivity. Sediment
anisotropy as well as scaling effects has also to be taken into
account.
Relationship between borehole lithology logs and AEM model resistivities Our analysis shows that resistivity generally increases
with grain size and permeability, as also shown by Burschil
et al. (2012b) and Klimke et al. (2013) for HTEM data. However,
there are substantial differences between the borehole lithology and resistivity values derived from the two interpolated
3D AEM resistivity grids (Fig. 11). The 3D HFEM resistivity grid
indicates that clay (grain-size class 1) has an average resistivity
of 30 m in the study area. This value corresponds to previous
results of AEM data commonly reported in the literature for clay
(Burschil et al., 2012b; Klimke et al., 2013). In comparison to
borehole resistivity (commonly between 5 and 20 m) the 3D
HFEM resistivity value is slightly to high and may be caused (1)
by a certain silt and sand content or (2) by the limited resolution of the HFEM method providing an over- or underestimated
thickness or merging of lithological units. The average resistivity value for deposits mainly consisting of clay to silt is 31 m
(grain-size class 2), for clay- and silt-rich fine sand is 31 m
(grain-size class 3), for diamicton is 80 m (grain-size class 4),
for fine sand is 105 m (grain-size class 5), for silt-rich fine to
coarse sand is 125 m (grain-size class 6) and for fine to coarse
sand with gravel is 125 m (grain-size class 7).
The relationship between borehole lithology and resistivity
data extracted from the grid based on 1D HTEM inversion results does not allow the lithology to be clearly defined in 3D

(Fig. 11). The average resistivity value for clay from our data
set (grain-size class 1) is 56 m and may be skewed due to the
small sample number. The average resistivity value for deposits
mainly consisting of clay to silt is 62 m (grain-size class 2),
for clay- and silt-rich fine sand is 53 m (grain-size class 3),
for diamicton is 115 m (grain-size class 4), for fine sand is
107 m (grain-size class 5), for silt-rich fine to coarse sand
is 55 m (grain-size class 6) and for fine to coarse sand with
gravel is 85 m (grain-size class 7).
The histograms for both resistivity data sets (see Fig. 11)
demonstrate that the measured values for sand are higher than
for deposits with certain clay content. However, overlapping
resistivity values for sand-dominated sediments and deposits
with clay content are generally found to have a lower resistivity
range, especially for the HTEM resistivity data set. The coarsergrained sediments (grain-size class 4 to 7) have a wide variety of
resistivity values and seem too low, especially for the HTEM data
set. Fig. 11 displays these large resistivity ranges. The variance
in values is interpreted to result from the mixed lithological
components, the limited resolution of the AEM system and the
less distinctive imaging of low-conductive sediments.
The larger overlaps of HTEM resistivity values can be explained as an effect of restricted HTEM data coverage and the
projection of interpolated resistivity values onto the borehole
locations. Similar HTEM resistivity values were found at the
island of Fohr (Burschil et al., 2012a,b) and for the region of
Quakenbruck,

southwestern Lower Saxony (Klimke et al., 2013).


Nonetheless, discrimination of different lithologies with a wide
range of resistivity values is possible by integrating with other
data sets, for example borehole or seismic reflection data.

Correlation of the HFEM resistivity model with the


geological model
Neogene marine and marginal marine deposits At depths between
10 and 120 m the spatial resistivity pattern is characterised
by an elongated structure with gradual large-scale variations
(100300 m) of medium to high resistivity values that indicate the heterogeneous infill of the Late Miocene incised valley
(seismic unit U4; Figs 4C and 5C). A positive correlation with
the hummocky seismic reflector pattern and borehole lithology
indicates that the resistivity pattern images grain-size variations of vertically and laterally stacked delta lobes (cf. Figs 4C
and 5C).
Pleistocene deposits The difference in lithology between the
fine-grained glacigenic Pleistocene deposits of the upper
tunnel-valley fill and the coarser-grained Neogene marginalmarine sediments is expressed as a distinct resistivity contrast,
which can be clearly traced in depth (Figs 4C, 5C and 6B).
The margin of the Cuxhaven tunnel valley is indicated by
a gradual shift from low to medium resistivity values of Pleistocene deposits (seismic subunits U5.5, U5.6 and U5.7; Figs 4C,

19

Netherlands Journal of Geosciences Geologie en Mijnbouw

5C, 6B and 7B) to higher resistivity values of the Pliocene host


sediments (seismic unit U4; Figs 4C, 5C, 6B and 7B). The resistivity pattern of the HFEM model enables a good 3D imaging
of the internal tunnel-valley fill. Large- and small-scale, elongated, wedge-shaped or lens-shaped variations in the resistivity
pattern can be correlated with major seismic units and smallerscale architectural elements, such as individual channels (cf.
Fig. 4C: U5.7 between 3800 and 4000 m) and lobes (cf. Fig. 4C:
U5.6 and U5.7 between 3600 and 3700 m).
The overall tabular geometry of the glaciolacustrine deposits
of the Lauenburg Clay Complex (seismic subunit U5.6) and the
marine Holsteinian interglacial deposits (seismic subunit U5.7)
are both characterised by low resistivity values in the tunnelvalley centre (Figs 4C, 5C, 6B and 7B) and gradually into higher
resistivity values towards the tunnel-valley margin. The higher
resistivity values towards the tunnel-valley margin have been
interpreted to be the result of bleeding of higher resistivity values from the neighbouring Pliocene sediments. Alternatively,
they could be interpreted as an indicator for coarse-grained material (e.g. delta foresets at the margin of the tunnel valley in
Fig. 5C: U5.2 and U5.4 between 700 and 800 m; in Fig. 4C: U5.5
at 4300 m and U5.6 at 4500 m). The comparison with lithology data from boreholes proves that the Lauenburg Clay Complex (seismic subunit U5.6) and the marine Holsteinian deposits
(seismic subunit U5.7) are approximately imaged at the correct
depths. The HFEM resistivity pattern, however, does not image a
distinct boundary between the glaciolacustrine Lauenburg Clay
Complex and the marine Holsteinian interglacial deposits due
to their similar grain sizes.
Within the tunnel-valley centre thick conductive finegrained beds of the Lauenburg Clay Complex limit the penetration depth of the HFEM system, leading to a decrease in
resolution and a less distinct resistivity pattern of underlying
fine- to medium-grained sand (cf. Fig. 5C, seismic subunit U5.5
in seismic line S2). Similar results have also been observed by
Steuer et al. (2009).
In the eastern part of the study area lower resistivity values in HFEM data are identified, which differ from the Neogene
host sediments. Its geometry and resistivity values suggest a
smaller-scale tunnel valley and probably indicate a fill of finegrained glaciolacustrine (Lauenburg Clay Complex) and/or interglacial marine (Holsteinian) deposits (Fig. 4C). The low resistivity contrast between this tunnel-valley fill and the Neogene
host sediments probably indicates relatively similar grain sizes,
but this remains speculative because no borehole log control is
available.
In the uppermost part of the Pleistocene succession between
0 and 10 m depth (seismic subunit U5.8; Figs 4C and 5C) a distinct shift from low to very high resistivity values indicates a
strong increase in electrical resistivity. This resistivity shift is
interpreted to represent the transition between water saturated
and unsaturated sediments it represents the groundwater table contact. This strong resistivity contrast allows the detection

20

of the groundwater table within a range of approximately 2 m


and clearly outlines the Hohe Lieth ridge as the most important
groundwater recharge area, which has also been documented by
Blindow & Balke (2005).

Correlation of the HTEM resistivity model with the


geological model
Neogene marginal-marine deposits At depths between 180 and
150 m (seismic unit U3) the resistivity pattern is characterised
by a sharp contrast from low to medium resistivity values
(Figs 4D, 5D, 6B and 7B). This upward increase in resistivity
is interpreted as an abrupt facies change from finer-grained
shelf deposits to coarser-grained shoreface deposits, which is
also recorded in borehole data (Kuster, 2005) and might be related to the rapid onset of progradation during the highstand
systems tract.
Correlation of the resistivity pattern with borehole data indicates that the low resistivity values link to marine Early Tortonian clays (seismic unit U3). At depths between 120 and
80 m, large-scale, lateral variations of low to medium resistivity
values, parallel to seismic reflections, can be identified. Altogether, resistivity contrast and seismic pattern, characterised
by strong reflectivity contrasts, are interpreted to result from
grain-size variations, which probably indicate the Tortonian
to Messinian storm-dominated shoreface deposits (seismic unit
U3; Table 1; Figs 4D and 5D; Walker & Plint, 1992; Catuneanu,
2002; Kuster, 2005; Catuneanu et al., 2011). Gradual vertical
transitions from low to high resistivity values within the upper Neogene unit (seismic unit U4; Figs 4D and 5D) correspond
to borehole lithology data with an overall coarsening-upward
trend are interpreted as prograding delta lobes (e.g. Dalrymple
et al., 1992; Plink-Bjorklund, 2008). This interpretation is further supported by the seismic pattern of lobate units (Figs 4A
and B and 5A and B).
Pleistocene deposits The lithology contrast between the finegrained glacigenic Pleistocene deposits at the base and the underlying coarser-grained Neogene marginal-marine sediments
is expressed as a distinct change in resistivity, which can be
clearly traced in the shallow subsurface between 20 and 30 m
depth and corresponds to the lower boundary of seismic unit
U5 (Figs 4D and 5D). The tunnel-valley margin can be identified in borehole data, but in the deeper subsurface is not clearly
imaged by the HTEM resistivity data. At the tunnel-valley margin the resistivity resembles that of the adjacent Neogene host
sediments and therefore its boundary may not be as easily
distinguished (Figs 4D, 5D, 6C and 7C). The inability for the
HTEM data to have contrasting resistivity values to define the
tunnel-valley boundary is probably caused by the large lateral
footprint of the HTEM system, the inversion approach (SCI) and
the kriging method, which leads to a smooth transition between
individual resistivity values.

Netherlands Journal of Geosciences Geologie en Mijnbouw

The upper Cuxhaven tunnel-valley fill is characterised by


different vertically stacked resistivity patterns. The highly conductive sediments of the Lauenburg Clay Complex lead to a reduced penetration of the EM field and the resistivity of the sediments below are not detectable with the transmitter moments
used in this survey. This suggests that the medium resistivity
values imaged at depths between 160 and 80 m do not necessarily show the true resistivities of Pleistocene sand within the
tunnel valley recorded in borehole data (seismic subunits U5.1
to U5.4; Figs 4D and 5D). However, the distinct shift towards
higher resistivity values in the unit below the highly conductive sediments of the Lauenburg Clay Complex probably indicates coarser-grained deposits of seismic subunit U5.5 (Figs 4D
and 5D). At depths between 80 and 10 m low resistivity values
can be correlated with fine-grained glaciolacustrine sediments
of the Lauenburg Clay Complex and interglacial Holsteinian deposits (seismic subunit U5.6 and U5.7; Figs 4D and 5D). The
comparison of the resistivity data with the borehole logs and
seismic data (Fig. 5B) indicates that resistivity changes reflect
lithological changes.

Verification of apparent resistivity values extracted


from the 3D subsurface models
The comparison of the apparent resistivity values derived from
the HFEM data (Fig. 12A), the initial geological subsurface
model based on borehole and seismic data (Fig. 12B), and the
adjusted geological subsurface model derived from the 3D AEM
resistivity grids (Fig. 12C) generally show a relationship indicating that both geological models are able to explain the
principal resistvity distribution at several HFEM frequencies.
The apparent resistivity images of the initial 3D geological
subsurface model show a relatively sharp resistivity contrast
between the Pleistocene tunnel valley and the adjacent Neogene deposits (Fig. 12B), particularly at the lower frequencies
(at greater depths) defining a simple tunnel-valley geometry
with a low sinuosity. The apparent resistivity map derived from
the initial 3D geological subsurface model (Fig. 12B), however,
does not image the tunnel-valley fill at the highest frequency
(at shallow depths). The difference between the apparent resistivity images can be explained by the limited coverage of
borehole and seismic data, which leads to restricted information about the subsurface architecture. The apparent resistivity
images of the adjusted 3D geological subsurface model derived
from the 3D AEM resistivity grids (Fig. 12C) show a more complex tunnel-valley geometry characterised by a higher sinuosity and smaller-scale variations of the resistivity pattern. This
more complex interpretation of the tunnel valley can be better aligned with the apparent resistivity images of HFEM data
(Fig. 12A) and the lithology variations recorded by borehole
data. Nevertheless, uncertainties remain and are caused by the
decreasing resolution with depth, which may lead to less contrast in the resistivity images. This is especially apparent at the

lowest frequency, whereby the resistivity images are influenced


by underlying sediments. This often leads to the bottom layer
of the 1D inversion models being represented by incorrect resistivity values. To reduce this problem we manually adjusted
the resistivity values of bottom units in the adjusted 3D geological subsurface model derived from the 3D AEM resistivity
grids based on lithology information (i.e. in particular the resistivity value for the lower part of unit U3 representing clay
and silt was reduced by a factor of about 2; cf. Table 1). The
resistivities of the valley infill were adjusted if the calculated
mean values were misleading for the sediment type (i.e. too
high resistivities attributed to clay and silt units were reduced,
e.g. seismic subunit U5.3, and too low resistivities attributed
to sandy units were increased, e.g. seismic subunits U5.1 and
U5.4; cf. Table 2). Some other values were rounded.
We re-calculated the apparent resistivity values after correction (Fig. 12D). These resistivity images better represent the
subsurface architecture and are closer to the apparent resistivity maps represented in the HFEM data, i.e. the adjusted
geological model better explains the HFEM data. Nevertheless,
differences in the images of the apparent resistivity patterns remain. This is caused by a too low resolution of the 3D geological
subsurface model, which does not represent detailed sediment
variability, and incorrectly estimated resistivities attributed to
the geological units. A good example of this effect is the area
of the smaller tunnel valley in the eastern part of the study
area (Fig. 12D), where the clay and silt deposits (lower part
of U3) are obviously thinner or/and resistive, and the central
part of the valley fill U5.5 was estimated as too broad and too
conductive. On the other hand, the low apparent resistivities
in the northwest of the study area, which occur particularly at
low frequencies, are caused by anthropogenic sources (airport
Cuxhaven/Nordholz).

Discussion
Our results show that AEM data provide excellent opportunities
to map the subsurface geology as previously demonstrated by,
for example, Newman et al. (1986), Jordan & Siemon (2002),
Danielsen et al. (2003), Jrgensen et al. (2003b, 2013), Auken &
Christiansen (2004), Auken et al. (2008), Viezzoli et al. (2008),
Bosch et al. (2009), Christensen et al. (2009), Klimke et al.
(2013) and Gunnink & Siemon (2014). A good relationship between resistivity values and lithology enables the 3D imaging
of the subsurface architecture. This allows a hitherto unseen
amount of geological detail using AEM data in areas of low
borehole coverage. This approach provides an advanced 3D geological model of the study area with new geological insights.
Although the presented approach is more time-consuming
than an automated approach that relies on statistics-based
methods (Bosch et al., 2009; Gunnink et al., 2012), the

21

Netherlands Journal of Geosciences Geologie en Mijnbouw

end results compensate for the limitations of the AEM


method.
Geological and geophysical models and their interpretations
contain limitations and uncertainties resulting from limitations
in the input information (Ross et al., 2005). Several studies focussed on the analysis of such limitations and uncertainties
in order to minimise risk, e.g. exploration risks (Bardossy &
Fodor, 2001; Pryet et al., 2011; Wellmann & Regenauer-Lieb,
2012; Jrgensen et al., 2013). Uncertainties and limitations in
the integrated interpretation of AEM and ground-based data
are mainly caused by the restricted availability and limited
vertical and lateral resolution of data (e.g. boreholes, seismic
sections, airborne surveys), modelling errors caused by a misinterpretation of geophysical, lithological and hydrogeological
properties, anthropogenic noise effects, missing software interoperability (Mann, 1993; Bardossy & Fodor, 2001; Ross et al.,
2005; Pryet et al., 2011) and irreducible, input-geology related,
uncertainty. In general there are several limitations in the interpretation caused by the applied AEM systems and the chosen
data analysis.
The penetration depth and resolution of both AEM systems
used is controlled by lithology and their conductivity. The
penetration depth is limited to approximately 100 m for HFEM
and 250 m for HTEM, and both systems are subject to decreasing
resolution capability with depth and require an increasing
thickness/depth ratio for the detection of varying depositional
units (Jrgensen et al., 2003b, 2005; Hyer et al., 2011). Because of the wide range of lithologies a significant uncertainty
remains in the final interpretation of the resistivity pattern.
Hence, thin-bedded sedimentary units are only resolved if their
corresponding conductance, i.e. the ratio of thickness and
resistivity, is sufficiently high. Otherwise they will be merged
into a single unit with an average resistivity (Jrgensen et al.,
2003b, 2005), which results in a limited resistivity resolution.
The limitations in lateral and vertical resolution can lead to
incorrect interpretations of thin-bedded sand/mud couplets,
smaller-scale architectural elements and bounding surfaces, as
has been also shown by, for example, Jrgensen et al. (2003a,
2005), Viezzoli et al. (2008), Christensen et al. (2009) and
Klimke et al. (2013). This problem is distinctly seen in our
study area in the Pleistocene tunnel valley and Late Miocene
incised valley. The Saalian till at 20 m depth with a thickness
between 5 and 10 m and a wide resistivity range (cf. Fig. 11)
could not be differentiated from the surrounding lithology
and is beyond the possible AEM resolution. The Pleistocene
tunnel-valley margins and the Late Miocene incised-valley
margin could not be detected because of a low resistivity
contrast to the adjacent Neogene host sediments.
For our study only an existing 1D AEM inversion data set
(from 2002) was available, in which noise effects were not
significantly minimised, as commonly applied for new data
sets (e.g. Tlbll, 2007; Siemon et al., 2011). However, the
noise effects of anthropogenic structures and soundings in the

22

airborne electromagnetic data can hinder a proper geological


interpretation, e.g. the airport Cuxhaven/Nordholz located in
the northwest of the study area. In the case of the study area
anthropogenic noise effects cause unrealistic low resistive oscillations, which conically increase downwards.
We used a geostatistical approach to analyse and interpolate the 1D AEM inversion results to create a continuous 3D
resistivity subsurface grid. We used ordinary kriging, and in
contrast to Pryet et al. (2011) we discarded the layered approach used in the 1D inversion of AEM data, leading to a
smoothing effect between previously defined layer boundaries.
The advantage of this method is that the kriging algorithm indicates the most likely resistivity value at each grid cell. The
result is an effectively smoothed resistivity grid, leading to
a significant improved geological interpretation of continuous
lithofacies. However, a loss of heterogeneity that is probably
observed in the subsurface by AEM has to be taken into account.
If the aim is to generate fluid-flow models it is important to
preserve the porosity and permeability of varying lithologies as
these strongly determine fluid-flow behaviour through porous
media (e.g. Koltermann & Gorelick, 1996; Janszen, 2012). It
is therefore preferred to base fluid-flow simulations on grids
that are interpolated using algorithms that are able to preserve
the heterogeneity, such as modified kriging methods like the
sequential indicator simulation used by Venteris (2007), Bosch
et al. (2009) and Janszen (2012).
A common limitation of the presented workflow is the high
degree of subjectivity that causes uncertainties that are hard to
evaluate. To test the reliability of the 3D geological subsurface
model and evaluate uncertainty, we followed a new approach to
derive synthetic HFEM data from the geological model results
based on a relationship of calculated or assumed geological
units and resistivities. The comparison between corresponding
apparent resistivity images of the subsurface based on AEM field
measurements and 3D geological models leads to the identification of uncertainties and limitations of 1D AEM inversion and
vice versa. The results of such analyses will offer the advantage of using a priori information to improve the inversion of
AEM data, comparable to investigations made by, for example,
Zhdanov (2010) and Gunnink et al. (2012).

Conclusions
Resistivity data derived from AEM surveys can provide a fast and
efficient overview mapping of the shallow subsurface geology. A
methodology for the interpretation, analysis and imaging of resistivity data gained by AEM surveys, based on 3D modelling approaches, was developed. We used 1D HFEM and HTEM inversion
results and applied a 3D resistivity gridding procedure based on
geostatistical analyses and interpolation techniques to create
continuous resistivity models of the subsurface. Subsequently,
we integrated results of seismic facies and borehole lithology

Netherlands Journal of Geosciences Geologie en Mijnbouw

analysis to construct a combined 3D geological subsurface model


and to reduce uncertainties. This approach allows for an improved interpretation of AEM data and imaging of the 3D subsurface architecture, if the relationship between lithology and
resistivity was known and the site was only slightly influenced
by noise effects (e.g. infrastructural networks, settlements).
The 3D resistivity models clearly allow discrimination between different lithologies and enable the detection of Cenozoic
sequence boundaries and larger-scale architectural elements
such as incised valleys and subglacial tunnel valleys. Variations in the resistivity pattern allowed the detection of individual delta lobes and smaller-scale channels as well as lateral
and vertical grain-size variations. In the Neogene succession
a low resistive pattern can be correlated with Tortonian and
Messinian marine clays. The lowest values probably correspond
with the zone of maximum flooding. From seismic interpretation we conclude that at the end of the Miocene an incised
valley formed as a response to a major sea-level fall that subsequently became filled with Pliocene-delta deposits, probably
indicating the palaeo-course of the Rivers Weser or Elbe. The
resistivity models clearly image the outline of the Pleistocene
tunnel valleys. The unconsolidated fill of the Late Miocene to
Pliocene incised valley probably formed a preferred pathway for
the Pleistocene meltwater flows, favouring the incision of a subglacial tunnel valley. Based on the 3D HFEM resistivity grid the
fills of the tunnel valleys could be imaged in much more detail
in comparison to previously published descriptions based on the
2D seismic sections and 2D AEM profiles and maps. The applied
approaches and results show a reliable methodology, especially
for future investigations of similar geological settings.

Auken, E. & Christiansen, A.V., 2004. Layered and laterally constrained 2D inversion of resistivity data. Geophysics 69(3): 752761.
Auken, E., Christiansen, A.V., Jacobsen, L.H. & Sorensen, K.I., 2008. A resolution study of buried valleys using laterally constrained inversion of TEM data.
Journal of Applied Geophysics 65(1): 1020.
Baldschuhn, R., Frisch, U. & Kockel, F. (eds), 1996. Geotektonischer Atlas
von NW-Deutschland 1 : 300000. Bundesanstalt fur
Geowissenschaften und
Rohstoffe (Hannover).
Baldschuhn, R., Binot, F., Fleig, S. & Kockel, F., 2001. Geotektonischer Atlas
von Nordwestdeutschland und dem deutschen Nordsee-Sektor. Geologisches
Jahrbuch A153: 395.
Bardossy, G. & Fodor, J., 2001. Traditional and new ways to handle uncertainty
in geology. Natural Resources Research 10(3): 179187.
Besenecker, H., 1976. Bohrdatenbank Niedersachsen, Bohr-ID: 3028HY0304, Landesamt fur
Bergbau, Energie und Geologie (LBEG) (Hannover).
Betz, D., Fuhrer, F., Greiner, G. & Plein, E., 1987. Evolution of the Lower Saxony
Basin. Tectonophysics 137: 127170.
Binot, F. & Wonik, T., 2005. Lithologie und Salz-/Suwassergrenze

in der
Forschungsbohrung Cuxhaven Ludingworth

1/1A. Zeitschrift fur


Angewandte
Geologie 1: 1423.
Blindow, N. & Balke, J., 2005. GPR-Messungen zur Bestimmung der Grundwasseroberflache im Bereich des Geestruckens

sudlich

von Cuxhaven. Zeitschrift fur

Angewandte Geologie 1: 3944.


Bosch, J.H.A., Bakker, M.A.J., Gunnink, J.L. & Paap, B.F., 2009. Airborne electromagnetic measurements as basis for a 3D geological model of an Elsterian
incision. Zeitschrift der deutschen Gesellschaft fur
Geowissenschaften 160(3):
249258.
Brandes, C., Pollok, L., Schmidt, C., Riegel, W., Wilde, V. & Winsemann, J.,
2012. Basin modelling of a lignite-bearing salt rim syncline: insights into
rim syncline evolution and salt diapirism in NW Germany. Basin Research 24:
699716.

Acknowledgements
The AIDA project is part of the joint research program of
GEOTECHNOLOGIEN. Financial support is gratefully acknowledged from the German Federal Ministry of Education and Research (BMBF) grant 03G0735. We thank two anonymous reviewers for very careful and constructive reviews, which helped
to improve the manuscript. Many thanks are due to Hajo Gotze
and Peter Menzel for discussion, and Guillaume Martelet for
helpful comments on an earlier version of the manuscript and
Ariana Osman for discussion and improving the English. We are
grateful to the State Authority for Mining, Energy and Geology
(LBEG) for providing borehole data, to Fugro for providing the
GeODin software and Paradigm for GOCAD software support.

Burschil, T., Wiederhold, H. & Auken, E., 2012a. Seismic results as a-priori
knowledge for airborne TEM data inversion A case study. Journal of Applied
Geophysics 80: 121128.
Burschil, T., Scheer, W., Kirsch, R. & Wiederhold, H., 2012b. Compiling geophysical and geological information into a 3-D model of the glacially-affected
island of Fohr. Hydrology and Earth System Sciences 16: 34853498.
BurVal Working Group, 2009. Buried Quaternary valleys - a geophysical approach.
Zeitschrift der deutschen Gesellschaft fur
Geowissenschaften 160(3): 237247.
Cameron, T.D.J., Buat, J. & Mesdag, C.S., 1993. High resolution seismic profile
through a Late Cenozoic delta complex in the southern North Sea. Marine and
Petroleum Geology 10: 591599.
Caspers, G., Jordan, H., Merkt, J., Meyer, K.-D., Muller, H. & Streif, H., 1995.
III. Niedersachsen. In: Benda, L. (ed.): Das Quartar Deutschlands. Gebruder

Borntrager (Berlin): 2358.


Catuneanu, O., 2002. Sequence stratigraphy of clastic systems: concepts, merits,
and pitfalls. Journal of African Earth Sciences 35: 143.

References

Catuneanu, O., Galloway, W.E., Kendall, C.G.St.C., Miall, A.D., Posamentier,


H.W., Strasser, A. & Tucker, M.E., 2011. Sequence Stratigraphy: Methodology

Anell, I., Thybo, H. & Rasmussen, E., 2012. A synthesis of Cenozoic sedimentation in the North Sea. Basin Research 24: 154179.
Archie, G.E., 1942. The electrical resistivity log as an aid in determining some
reservoir characteristics. Trans AIME 146: 5462.

and Nomenclature. Newsletters on Stratigraphy 44(3): 173245.


Christensen, N.B., Reid, J.E. & Halkjr, M., 2009. Fast, laterally smooth inversion
of airborne time-domain electromagnetic data. Near Surface Geophysics 7:
599612.

23

Netherlands Journal of Geosciences Geologie en Mijnbouw

Dalrymple, R.W., Zaitlin, B.A. & Boyd, R., 1992. Estuarine facies models: conceptual basis and stratigraphic implications. Journal of Sedimentary Petrology
62: 11301146.

tection. Near Surface Geophysics, 12 (early online), doi: 10.3997/18730604.2014044. EAGE Publications (Houten).
Gunnink, J.L., Bosch, J.H.A., Siemon, B., Roth, B. & Auken, E., 2012. Com-

Danielsen, J.E., Auken, E., Jrgensen, F., Sndergard, V. & Srensen, K.I., 2003.

bining ground-based and airborne EM through artificial neural networks for

The application of the transient electromagnetic method in hydrogeological

modelling hydrogeological units under saline groundwater conditions. Hy-

surveys. Journal of Applied Geophysics 53: 181198.

drology and Earth System Sciences 16: 30613074.

de Louw, P.G.B., Eeman, S., Siemon, B., Voortman, B.R., Gunnink, J., van
Baaren, E.S. & Oude Essink, G.H.P., 2011. Shallow rainwater lenses in deltaic
areas with saline seepage. Hydrology Earth System Sciences 15: 36593678.
Ehlers, J., 2011. Geologische Karte von Hamburg 1: 25 000: Erlauterungen zu
Blatt Nr. 2326 Fuhlsbuttel.

Behorde fur
Stadtentwicklung und Umwelt, Geologisches Landesamt Hamburg (Hamburg).

Haq, B.U., Hardenbol, J. & Vail, P.R., 1987. Chronology of fluctuating sea levels
since the Triassic (250 million years ago to present). Science 235: 11561167.
Hese, F., 2012. 3D Modellierungen und Visualisierung von Untergrundstrukturen
fur
die Nutzung des unterirdischen Raumes in Schleswig-Holstein. PhD thesis,
Christian-Albrechts-Universitat zu Kiel (Kiel).
HGG, 2011. Guide to processing and inversion of SkyTEM data in the Aarhus

Ehlers, J. & Linke, G., 1989. The origin of deep buried channels of Elsterian age
in northwest Germany. Journal of Quaternary Science 4: 255265.
Ehlers, J., Meyer, K.-D. & Stephan, H.-J., 1984. Pre-weichselian glaciations of
north-west Europe. Quaternary Science Reviews 3: 140.

Workbench. Technical report, Hydro-Geophysical Group, Department of Earth


Sciences, University of Aarhus (Aarhus).
Hofle, H.C., Merkt, J. & Muller, H., 1985. Die Ausbreitung des Eem-Meeres in
Nordwestdeutschland. Eiszeitalter und Gegenwart 35: 4959.

Ehlers, J., Grube, A., Stephan, H.-J. & Wansa, S., 2011. Pleistocene Glaciations

Hyer, A.-S., Lykke-Andersen, H., Jrgensen, F. & Auken, E., 2011. Combined

of North Germany New Results. In: Ehlers, J., Gibbard, P.L. & Hughes,

interpretation of SkyTEM and high-resolution seismic data. Journal of Physics

P.D. (eds): Quaternary Glaciations Extent and Chronology A Closer Look.


Developments in Quaternary Science 15. Elsevier (Amsterdam): 149162.
Eidvin, T. & Rundberg, Y, 2007. Post-Eocene strata of the southern Viking
Graben, northern North Sea; integrated biostratigraphic, strontium isotopic
and lithostratigraphic study. Norwegian Journal of Geology 87: 391450.
Fraser, D.C., 1978. Resistivity mapping with an airborne multicoil electromagnetic
system. Geophysics 43: 144172.

and Chemistry of the Earth 36(16): 13861397.


Huuse, M., 2002. Late Cenozoic palaeogeography of the eastern North Sea Basin
climatic vs. tectonic forcing of basin margin uplift and deltaic progradation.
Bullettin of the Geological Society of Denmark 49: 145170.
Huuse, M. & Clausen, O.R., 2001. Morphology and origin of major Cenozoic
sequence boundaries in the eastern North Sea Basin: top Eocene, near-top
Oligocene and the mid-Miocene unconformity. Basin Research 13: 1741.

Fugro Consult GmbH, 2012. GeODin 7, http://www.geodin.com. Berlin, Germany.

Huuse, M. & Lykke-Andersen, H., 2000. Overdeepened Quaternary valleys in the

Gabriel, G., 2006. Gravity investigation of buried Pleistocene subglacial valleys.

eastern Danish North Sea: morphology and origin. Quaternary Science Reviews

Near Surface Geophysics 4: 321332.

19: 12331253.

Gabriel, G., Kirsch, R., Siemon, B. & Wiederhold, H., 2003. Geophysical investi-

Huuse, M., Lykke-Andersen, H. & Michelsen, O., 2001. Cenozoic evolution of the

gations of buried Pleistocene subglacial valleys in Northern Germany. Journal

eastern North Sea Basin new evidence from high-resolution and conven-

of Applied Geophysics 53: 159180.

tional seismic data. Marine Geology 177: 243269.

Gast, R. & Gundlach, T., 2006. Permian strike slip and extensional tectonics in

Huuse, M., Piotrowski, J.A. & Lykke-Andersen, H., 2003. Geophysical investi-

Lower Saxony, Germany. Zeitschrift der Deutschen Gesellschaft fur


Geowis-

gations of buried Quaternary valleys in the formerly glaciated NW European

senschaften 157(1): 4155.

lowland: significance for groundwater exploration. Journal of Applied Geo-

Gibbard, P.L., 1988. The history of the great northwest European rivers during
the past three million years. Philosophical Transactions of the Royal Society
of London B318: 559602.

tions. Geologisches Jahrbuch A100: 410423.


F.,

1989.

Forschungsbohrung

Janszen, A., Spaak, M. & Moscariello, A., 2012. Effects of the substratum on the
formation of glacial tunnel valleys: an example from the Middle Pleistocene

Wursterheide.

Benthonische

Foraminiferen und verwandte Mikrofossilien. Biostratigraphie, Faziesanalyse.


Geologisches Jahrbuch A111: 287319.
Gramann, F. & Daniels, C.H. von, 1988. Benthic foraminifera: the description of
the interregional zonation (B zones). Geologisches Jahrbuch A100: 145151.
Gramann, F. & Kockel, F., 1988. Palaeogeographical, lithological, palaeoecological and palaeoclimatic development of the Northwest European Tertiary Basin.
Geologisches Jahrbuch A100: 428441.
Grassmann, S., Cramer, B., Delisle, G., Messner, J. & Winsemann, J., 2005.
Geological history and petroleum system of the Mittelplate oil field, Northern
Germany. International Journal of Earth Sciences 94: 979989.
Gunnink, J.L. & Siemon, B., 2014. Applying airborne electromagnetics in
3D stochastic geohydrological modelling for determining groundwater pro-

24

Janszen, A., 2012. Tunnel valleys: genetic models, sedimentary infill and 3D
architecture. PhD thesis, Delft University of Technology (Delft).

Gramann, F., 1988. Major palaeontological events and biostratigraphical correla-

Gramann,

physics 53: 153157.

of the southern North Sea Basin. Boreas 41: 629643.


Janszen, A., Moreau, J., Moscariello, A., Ehlers, J. & Kroger, J., 2013. Timetransgressive tunnel-valley infill revealed by a three-dimensional sedimentary
model, Hamburg, north-west Germany. Sedimentology 60(3): 693719.
Jaritz, W., 1987. The origin and development of salt structures in Northwest
Germany. In: Lerche, I. & OBrian, J.J. (eds): Dynamical Geology of Salt and
Related Structures. Academic Press (Orlando): 479493.
Jordan, H. & Siemon, B., 2002. Die Tektonik des nordwestlichen Harzrandes
Ergebnisse der Hubschrauber-Elektromagnetik. Zeitschrift der Deutschen
Geologischen Gesellschaft 153(1): 3150.
Jrgensen, F. & Sandersen, P.B.E., 2006. Buried and open tunnel valleys in
Denmark erosion beneath multiple ice sheets. Quaternary Science Reviews
25: 13391363.

Netherlands Journal of Geosciences Geologie en Mijnbouw

Jrgensen, F. & Sandersen, P.B.E., 2009. Buried valley mapping in Denmark:

Kothe, A., 2007. Cenozoic biostratigraphy from the German North Sea sector (G-

evaluating mapping method constraints and the importance of data density.

11-1 borehole, dinoflagellate cysts, calcareous nannoplankton). Zeitschrift

Zeitschrift der deutschen Gesellschaft fur


Geowissenschaften 160(3): 211229.

der deutschen Gesellschaft fur


Geowissenschaften 158(2): 287327.

Jrgensen, F., Lykke-Andersen, H., Sandersen, P.B.E., Auken, E. & Nrmark, E.,

Kothe, A., Gaedicke, C. & Rudiger, L., 2008. Erratum: The age of the Mid-

2003a. Geophysical investigations of buried Quaternary valleys in Denmark:

Miocene Unconformity (MMU) in the G-11-1 borehole, German North Sea

an integrated application of transient electromagnetic soundings, reflection

sector. Zeitschrift der deutschen Gesellschaft fur


Geowissenschaften 159(4):

seismic surveys and exploratory drillings. Journal of Applied Geophysics 53:


215228.
Jrgensen, F., Sandersen, P.B.P. & Auken, E., 2003b. Imaging buried valleys
using the transient electromagnetic method. Journal of Applied Geophysics
53: 215228.

687689.
Krige, D.G., 1951. A statistical approach to some basic mine valuation problems
on the Witwatersrand. Journal of the Chemical, Metallurgical and Mining
Society of South Africa 52(6): 119139.
Kuhlmann, G., de Boer, P., Pedersen, R.B. & Wong, Th.E., 2004. Provenance of

Jrgensen, F., Sandersen, P.B.P., Auken, E., Lykke-Andersen, H. & Srensen,

Pliocene sediments and paleoenvironmental changes in the southern North

K., 2005. Contributions to the geological mapping of Mors, Denmark A

Sea region using Samarium-Neodymium (Sm/Nd) provenance ages and clay

study based on large-scale TEM survey. Bulletin of the Geological Society of

mineralogy. Sedimentary Geology 171: 205226.

Denmark 52: 5375.


Jrgensen, F., Rnde Mller, R., Nebel, L., Jensen, N.-P., Christiansen, A.V. &
Sandersen, P.B.E., 2013. A method for cognitive 3D geological voxel modelling of AEM data. Bulletin of Engineering Geology and the Environment
72(34): 421432.
Jurgens, U., 1996. Mittelmiozane bis pliozane Randmeer-Sequenzen aus dem
deutschen Sektor der Nordsee. Geologisches Jahrbuch A146: 217232.
Kehew, A.E., Piotrowski, J.A. & Jrgensen, F., 2012. Tunnel valleys: Concepts
and controversies A review. Earth-Science Reviews 113: 3358.
Klimke, J., Wiederhold, H., Winsemann, J., Ertl, G. & Elbracht, J., 2013.

Kuhlmann, G. & Wong, T.E., 2008. Pliocene paleoenvironment evolution as interpreted from 3D-seismic data in the southern North Sea, Dutch offshore
sector. Marine and Petroleum Geology 25: 173189.
Kuster, H., 2005. Das jungere

Tertiar in Nord- und Nordostniedersachsen. Geologisches Jahrbuch A158: 0194.


Kuster, H. & Meyer, K.-D., 1979. Glaziare Rinnen im mittleren und nordostlichen
Niedersachsen. Eiszeitalter und Gegenwart 29: 135156.

Kuster, H. & Meyer, K.-D., 1995. Quartargeologische Ubersichtskarte


von Niedersachsen und Bremen, 1 : 500 000. Niedersachsisches Landesamt fur
Bodenforschung (Hannover).

Three-dimensional mapping of Quaternary sediments improved by airborne

Lane, R., Green, A., Golding, C., Owers, M., Pik, P., Plunkett, C., Sattel, D. &

electromagnetics in the case of the Quakenbruck


Basin, Northern Germany.

Thorn, B., 2000. An example of 3D conductivity mapping using the TEMPEST

Zeitschrift der deutschen Gesellschaft fur


Geowissenschaften 164(2): 369384.

airborne electromagnetic system. Exploration Geophysics 31: 162172.

Kluiving, S.J., Bosch, J.H.A., Ebbing, J.H.J., Mesdag, C.S. & Westerhoff, R.S.,

Lang, J., Winsemann, J., Steinmetz, D., Polom, U., Pollok, L., Bohner, U.,

2003. Onshore and offshore seismic and lithostratigraphic analysis of a deeply

Serangeli, J., Brandes, C., Hampel, A. & Winghart, S., 2012. The Pleis-

incised Quaternary buried valley-system in the Northern Netherlands. Journal

tocene of Schoningen, Germany: a complex tunnel valley fill revealed from

of Applied Geophysics 53: 249271.

3D subsurface modelling and shear wave seismics. Quaternary Science Reviews

Knox, R.W.O.B., Bosch, J.H.A., Rasmussen, E.S., Heilmann-Clausen, C., Hiss,

39: 86105.

M., De Lugt, I.R., Kasinksi, J., King, C., Kothe, A., Sodkowska, B., Standke,

Linke, G., 1993. Zur Geologie und Petrographie der Forschungsbohrungen qho

G. & Vandenberghe, N., 2010. Cenozoic. In: Doornenbal, J.C. & Stevenson,

15, der Bohrung Hamburg-Billbrook und des Vorkommens von marinem Hol-

A.G. (eds): Petroleum Geological Atlas of the Southern Permian Basin Area.

stein im Gebiet Neuwerk-Scharhorn. Geologisches Jahrbuch A138: 3576.

EAGE Publications b.v. (Houten): 211223.


Knudsen, K.L., 1988. Marine Interglacial Deposits in the Cuxhaven Area, NW
Germany: A Comparison of Holsteinian, Eemian and Holocene Foraminiferal
Faunas. Eiszeitalter und Gegenwart 38: 6977.
Knudsen, K.L., 1993a. Foraminiferal Faunas in Holsteinian Deposits of the Neuwerk Area, Germany. Geologisches Jahrbuch A138: 7795.
Knudsen, K.L., 1993b. Late Elsterian-Holsteinian Foraminiferal Stratigraphy in

Litt, T., Behre, K.-E., Meyer, K.-D., Stephan, H.-J. & Wansa, S., 2007. Stratigraphische Begriffe fur
das Quartar des norddeutschen Vereisungsgebietes.
Eiszeitalter und Gegenwart 56: 765.
Lutz, R., Kalka, S., Gaedicke, C., Reinhardt, L. & Winsemann, J., 2009. Pleistocene tunnel valleys in the German North Sea: spatial distribution and morphology. Zeitschrift der deutschen Gesellschaft fur
Geowissenschaften 160(3):
225235.

Boreholes in the Lower Elbe Area, NW Germany. Geologisches Jahrbuch A138:

Mallet, J.-L., 2002. Geomodeling. Oxford University Press (New York): 624 pp.

97119.

Mangerud, J., Jansen, E. & Landvik, J.Y., 1996. Late Cenozoic history of the

Kockel, F., 2002. Rifting processes in NW-Germany and the German North Sea
Sector. Netherlands Journal of Geosciences/Geologie en Mijnbouw 81: 149
158.
Koltermann, C.E. & Gorelick, S.M., 1996. Heterogeneity in sedimentary deposits: a review of structure-imitating, process-imitating, and descriptive
approaches. Water Resources Research 32: 26172658.
Konradi, P., 2005. Cenozoic stratigraphy in the Danish North Sea Basin. Geologie
en Mijnbouw 84: 109111.

Scandinavian and Barents Sea ice sheets. Global and Planetary Change 12:
1126.
Mann, C.J., 1993. Uncertainty in geology. In: Davis, J.C. & Herzfeld, U.C. (eds):
Computers in geology 25 years of progress. Oxford University Press (Oxford):
241254.
Maystrenko, Y., Bayer, U. & Scheck-Wenderoth, M., 2005a. The Glueckstadt
Graben, a sedimentary record between the North- and Baltic Sea in North
Central Europe. Tectonophysics 397: 113126.

25

Netherlands Journal of Geosciences Geologie en Mijnbouw

Maystrenko, Y., Bayer, U. & Scheck-Wenderoth, M., 2005b. Structure and evolu-

Pharaoh, T.C., Dusar, , Geluk, M., Kockel, M.C., Krawczyk, F., Krzywiec, C.M.,

tion of the Glueckstadt Graben due to salt movements. International Journal

Scheck-Wenderoth, P., Thybo, M., Vejbk, H., & Wees, O.V., van, J.D., 2010.

of Earth Science 94: 799814.

Tectonic evolution. In: Doornenbal, J.C. & Stevenson, A.G., (eds): Petroleum

Michelsen, O., Danielsen, M., Heilmann-Clausen, C., Jordt, H., Laursen, G.V.
& Thomsen, E., 1995. Occurrence of major sequence stratigraphic boundaries

Geological Atlas of the Southern Permian Basin Area. EAGE Publications b.v.
(Houten): 2557.

in relation to basin development in Cenozoic deposits of the southeastern

Plink-Bjorklund, P., 2008. Wave-to-tide facies change in a Campanian shoreline

North Sea. In: Steel, R.J., Felt, V.L., Johannesen, E.P. & Mathieu, C. (eds):

complex, Chimney Rock Tongue, Wyoming-Utah, U.S.A. In: Hampson, G.J.,

Sequence Stratigraphy of the Northwest European Margin. Special Publications

Steel, R.J., Burgess, P.M. & Dalrymple, R.W. (eds): Recent Advances in Models

5, Norwegian Petroleum Society (Stavanger, Norway): 415427.

of Siliciclastic Shallow-Marine Stratigraphy. Special Publications 90, Society

Michelsen, O., Thomsen, E., Danielsen, M., Heilmann-Clausen, C., Jordt, H. &

for Sedimentary Geology (Oklahoma): 265292.

Laursen, G.V., 1998. Cenozoic sequence stratigraphy in the eastern North

Praeg, D., 2003. Seismic imaging of mid-Pleistocene tunnel-valleys in the North

Sea. In: Graciansky, P.-C. de, Hardenbol, J., Jacquin, T. & Vail, P.R. (eds):

Sea Basin high resolution from low frequencies. Journal of Applied Geo-

Mesozoic and Cenozoic Sequence Stratigraphy of European Basins. Special


Publication 60, Society for Sedimentary Geology (Oklahoma): 91118.
Miller, K.G., Kominz, M.A., Browning, J.V., Wright, J.D., Mountain, G.S.,
Katz, M.E., Sugarman, P.J., Cramer, B.S., Christie-Blick, N. & Pekar, S.F.,
2005. The Phanerozoic record of global sea-level change. Science 310: 1293
1298.
Mitchum, R.M., Vail, & Sangree, P.R., 1977, J.B. Seismic stratigraphy and global
changes of sea-level, Part 6: stratigraphic interpretation of seismic reflection
patterns in depositional sequences. In: Payton, C.E. (ed.): Seismic Stratigraphy Applications to Hydrocarbon Exploration. AAPG Memoir 26 (Houston):
117133.
Mller, L.K., Rasmussen, E.S. & Clausen, O. R., 2009. Clinoform migration patterns of a Late Miocene delta complex in the Danish Central Graben; implications for relative sea-level changes. Basin Research 21: 704720.
Moreau, J., Huuse, M., Janszen, A., van der Vegt, P., Gibbard, P.L. &

physics 53: 273298.


Pryet, A., Ramm, J., Chil`es, J.-P., Auken, E., Deffontaines, B. & Violette, S.,
2011. 3D resistivity gridding of large AEM datasets: A step toward enhanced
geological interpretation. Journal of Applied Geophysics 75: 277283.
Rasmussen, E.S., 2004. The interplay between true eustatic sea-level changes,
tectonics, and climatic changes: what is the dominating factor in sequence
formation of the Upper OligoceneMiocene succession in the eastern North
Sea Basin, Denmark? Global and Planetary Change 41: 1530.
Rasmussen, E.S. & Dybkjr, K., 2013. Patterns of Cenozoic sediment flux from
western Scandinavia: discussion. Basin Research 25: 19.
Rasmussen, E.S., Heilmann-Clausen, C., Waagstein, R. & Eidvin, T., 2008. The
Tertiary of Norden. Episodes 31: 6672.
Rasmussen, E.S., Dybkjr, K. & Piasecki, S., 2010. Lithostratigraphy of the Upper
Oligocene Miocene succession of Denmark. Geological Survey of Denmark
and Greenland Bulletin 22: 92 pp.

Moscariello, A., 2012. The glaciogenic unconformity of the southern North

Roskosch, J., Winsemann, J., Polom, U., Brandes, C., Tsukamoto, S., Weitkamp,

Sea. In: Huuse, M., Redfern, J., Le Heron, D.P., Dixon, R.J., Moscariello, A.

A., Bartholomaus, W.A., Henningsen, D. & Frechen, M., 2014. Luminescence

& Craig, J. (eds): Glaciogenic Reservoirs and Hydrocarbon Systems. Special

dating of ice-marginal deposits in northern Germany: evidence for repeated

Publications 368, Geological Society (London): 99110.

glaciations during the Middle Pleistocene (MIS 12 to MIS 6). Boreas. DOI

Muller, H. & Hofle, H.-C., 1994. Die Holstein-Interglazialvorkommen bei Bossel


westlich von Stade und Wanhoden nordlich Bremerhaven. Geologisches
Jahrbuch A134: 71116.
Newman, G.A., Hohmann, G.W. & Anderson, W.L., 1986. Transient electromagnetic response of a 3-dimensional body in layered earth. Geophysics 51(8):
16081627.
Odin, G.S. & Kreuzer, H., 1988. Geochronology: some geochronological calibration
points for lithostratigraphic units. Geologisches Jahrbuch A100: 403410.
Ortlam, D., 2001. Geowissenschaftliche Erkenntnisse uber

den Untergrund Bremerhavens. Bremisches Jahrbuch 80: 181197.

10.1111/bor.12083. Wiley-Blackwell (Hoboken).


Ross, M., Parent, M. & Lefebvre, R., 2005. 3D geologic framework models for
regional hydrogeology and land-use management: a case study from a Quaternary basin of southwestern Quebec, Canada. Hydrogeology Journal 13:
690707.
Rumpel, H.-M., Grelle, T., Gruneberg, S., Gromann, E. & Rode, W., 2006a. Reflexionsseismische Untersuchungen der Cuxhavener Rinne bei Midlum 2005.
Technischer Bericht, GGA-Institut, Archiv-Nr. 126 015 (Hannover).
Rumpel, H.-M., Binot, F., Gabriel, G., Hinsby, K., Siemon, B., Steuer, A. &
Wiederhold, H., 2006b. Cuxhavener Rinne. In: Kirsch, R., Rumpel, H.-M.,

Overeem, I., Weltje, G.J., Bishop-Kay, C. & Kroonenberg, S.B., 2001. The late

Scheer, W. & Wiederhold, H. (eds): BurVal Working Group: Groundwater re-

Cenozoic Eridanos delta system in the southern North Sea Basin: a climate

sources in buried valleys a challenge for geosciences. Bonifatius GmbH

signal in sediment supply? Basin Research 13: 293312.

(Hannover): 227240.

Paine, J.G. & Minty, B.R.S., 2005. Airborne hydrogeophysics. In: Rubin, Y. &
Hubbard, S.S. (eds): Hydrogeophysics. Springer (Dordrecht): 333357.
Palamara, D.R., Rodriguez, V.B., Kellett, J. & Macaulay, S., 2010. Salt mapping
in the Lower Macquarie area, Australia, using airborne electromagnetic data.
Environmental Earth Sciences 61: 613623.
Panteleit, B. & Hammerich, T., 2005. Hydrochemische Charakteristiken und
Prozesse im Kustenbereich

bei Cuxhaven. Zeitschrift fur


Angewandte Geologie
1: 6067.
R
R
, 2011. GOCAD
. Version 2009.3, Patch 2.
Paradigm

26

Rumpel, H.-M., Binot, F., Gabriel, G., Siemon, B., Steuer, A. & Wiederhold,
H., 2009. The benefit of geophysical data for hydrogeological 3D modelling
- an example using the Cuxhaven buried valley. Zeitschrift der deutschen
Gesellschaft fur
Geowissenschaften 160(3): 259269.
Rundberg, Y. & Smalley, P.C., 1989. High-resolution dating of Cenozoic sediments from the northern North Sea using

87

Sr/86 Sr stratigraphy. AAPG Bul-

letin 73: 298308.


Russell, H.A.J., Brennand, T.A., Logan, C. & Sharpe, D.R., 1998. Standardization
and assessment of geological descriptions from water well records, Greater

Netherlands Journal of Geosciences Geologie en Mijnbouw

Toronto and Oak Ridge Moraine areas, southern Ontario. In: Current Research,
part E, Geological Survey of Canada, Paper 1998-E: 89102.
Sandersen, P.B.E. & Jrgensen, F., 2003. Buried Quaternary valleys in western
Denmark occurrence and inferred implications for groundwater resources
and vulnerability. Journal of Applied Geophysics 53: 229248.
Scheck-Wenderoth, M. & Lamarche, J., 2005. Crustal memory and basin evolution
in the Central European Basin System-new insights from a 3D structural model.
Tectonophysics 397: 143165.
Sengpiel, K.-P. & Siemon, B., 2000. Advanced inversion methods for airborne
electromagnetics. Geophysics 65: 19831992.
Siemon, B., 2001. Improved and new resistivity-depth profiles for helicopter
electromagnetic data. Journal of Applied Geophysics 46: 6576.
Siemon, B., 2005. Ergebnisse der Aeroelektromagnetik zur Grundwassererkundung im Raum Cuxhaven-Bremerhaven. Zeitschrift fur
Angewandte Geologie
1: 713.

Streif, H., 2004. Sedimentary record of Pleistocene and Holocene marine inundations along the North Sea coast of Lower Saxony, Germany. Quaternary
International 112(1): 328.
Tezkan, B., Mbiyah, M.H., Helwig, S.L. & Bergers, R., 2009. Time domain electromagnetic (TEM) measurements on a buried subglacial valley in Northern
Germany by using a large transmitter size and a high current. Zeitschrift der
deutschen Gesellschaft fur
Geowissenschaften 160(3): 271278.
Thole, H., Gaedicke, C., Kuhlmann, G. & Reinhardt, L., 2014. Late Cenozoic
sedimentary evolution of the German North Sea A seismic stratigraphic
approach. Newsletters on Sratigraphy 47(3): 299329.
Tlbll, R.J., 2007. The application of frequency-domain helicopter-borne electromagnetic methods to hydrogeological investigations in Denmark. PhD thesis,
Department of Earth Sciences, University of Aarhus (Aarhus).
van der Vegt, P., Janszen, A. & Moscariello, A., 2012. Glacial tunnel valleys
current knowledge and future perspectives. In: Huuse, M., Redfern, J.,

Siemon, B., Stuntebeck, C., Sengpiel, K.-P., Rottger, B., Rehli, H.-J. & Eberle,

Le Heron, D.P., Dixon, R.J., Moscariello, A. & Craig, J. (eds): Glaciogenic

D., 2002. Characterisation of hazardous waste sites using the BGR helicopter-

Reservoirs and Hydrocarbon Systems, Special Publications 368, Geological

borne geophysical system. Journal of Environmental and Engineering Geo-

Society (London): 7597.

physics 7: 169181.
Siemon, B., Eberle, D.G. & Binot, F., 2004. Helicopter-borne electromagnetic
investigation of coastal aquifers in North-West Germany. Zeitschrift fur
geologische Wissenschaften 32: 385395.
Siemon, B., Christiansen, A.V. & Auken, E., 2009a. A review of helicopterborne electromagnetic methods for groundwater exploration. Near Surface
Geophysics 7: 629646.

Venteris, E.R., 2007. Three-dimensional modeling of glacial sediments using public water-well data records: an integration of interpretive and geostatistical
approaches. Geospheres 3: 456468.
Viezzoli, A., Christiansen, A.V., Auken, E. & Srensen, K., 2008. Quasi-3D modeling of airborne TEM data by spatially constrained inversion. Geophysics
73(3): F105F113.
Walker, R.G. & Plint, A.G., 1992. Wave and storm-dominated shallow marine

Siemon, B., Auken, E. & Christiansen, A.V., 2009b. Laterally constrained inver-

systems. In: Walker, R.G. & James, N.P. (eds): Facies Models: Response

sion of helicopter-borne frequency-domain electromagnetic data. Journal of

to Sea Level Change. Geological Association of Canada (St Johns): 219

Applied Geophysics 67: 259268.

238.

Siemon, B., Steuer, A., Ullmann, A., Vasterling, M. & Vo, M., 2011. Applica-

Wellmann, J.F. & Regenauer-Lieb, K., 2012. Effect of geological data qual-

tion of frequency-domain helicopter-borne electromagnetics for groundwater

ity on uncertainties in geological model and subsurface flow fields. PRO-

exploration in urban areas. Physics and Chemistry of the Earth 36: 13731385.

CEEDINGS, Thirty-Seventh Workshop on Geothermal Reservoir Engineer-

Srensen, K. & Auken, E., 2004. SkyTEM - a new high-resolution helicopter

ing Stanford University, Stanford, California, January 30February 1, 2012

transient electromagnetic system. Exploration Geophysics 35: 191199.

SGP-TR-194.

Srensen, J.C., Gregersen, U., Breiner, M. & Michelsen, O., 1997. High-frequency

West, G.F. & Macnae, J.C., 1991. Physics of the electromagnetic induction ex-

sequence stratigraphy of Upper Cenozoic deposits in the central and south-

ploration method. In: Nabighian, M.N. & Corbett, J.D. (eds): Electromagnetic

eastern North Sea areas. Marine and Petroleum Geology 14(2): 99123.

methods in applied geophysics. Investigations in geophysics, Society of Ex-

Stackebrandt, W., 2009. Subglacial channels of Northern Germany a brief review. Zeitschrift der deutschen Gesellschaft fur
Geowissenschaften 160(3):
203210.
Steuer, A., 2008. Joint application of ground based transient electromagnetics
and airborne electromagnetics. PhD thesis, University of Cologne (Cologne).
Steuer, A., Siemon, B. & Auken, E., 2009. A comparison of helicopter-borne
electromagnetics in frequency- and time-domain at the Cuxhaven valley in
Northern Germany. Journal of Applied Geophysics 67: 194205.
Stoker, M.S., Hoult, R.J., Nielsen, T., Hjelstuen, B.O., Laberg, J.S., Shannon,
P.M., Praeg, D., Mathiesen, A., Weering, T.C.E. van & McDonnell, A., 2005a.
Sedimentary and oceanographic responses to early Neogene compression on
the NW European margin. Marine Petroleum Geology 22: 10311044.
Stoker, M.S., Praeg, D., Hjelstuen, B.O., Laberg, J.S., Nielsen, T. & Shannon,
P.M., 2005b. Neogene stratigraphy and the sedimentary and oceanographic
development of the NW European Atlantic margin. Marine Petroleum Geology
22: 9771005.

ploration Geophysicists 2: 545.


Wiederhold, H., Binot, F. & Kessels, W., 2005a. Die Forschungsbohrung Cuxhaven
und das ,,Coastal Aquifer Testfield (CATField) ein Testfeld fur
angewandte
geowissenschaftliche Forschung. Zeitschrift fur
Angewandte Geologie 1: 35.
Wiederhold, H., Gabriel, G. & Grinat, M., 2005b. Geophysikalische Erkundung der
Bremerhaven-Cuxhavener Rinne im Umfeld der Forschungsbohrung Cuxhaven.
Zeitschrift fur
Angewandte Geologie 1: 2838.
Wingfield, R., 1990. The origin of major incisions within the Pleistocene deposits
of the North Sea. Marine Geology 91: 3152.
2001. Seismic data analysis: processing, inversion and interpretation
Yilmaz, O.,
of seismic data. Society of Exploration Geophysicists (Tulsa): 2027 pp.
Zachos, J., Pagani, M., Sloan, L., Thomas, E. & Billups, K., 2001. Trends,
rhythms, and aberrations in global climate 65 Ma to present. Science 292:
686693.
Zhdanov, M.S., 2010. Electromagnetic geophysics: notes from the past and the
road ahead. Geophysics 75(5): 75A4975A66.

27